Population-Scale Study of Human Needs
During the COVID-19 Pandemic: Analysis and Implications

Jina Suh [email protected] University of Washington Microsoft Research , Eric Horvitz, Ryen W. White horvitz,[email protected] Microsoft Research and Tim Althoff [email protected] University of Washington

Abstract.

Most work to date on mitigating the COVID-19 pandemic is focused urgently on biomedicine and epidemiology. Yet, pandemic-related policy decisions cannot be made on health information alone. Decisions need to consider the broader impacts on people and their needs. Quantifying human needs across the population is challenging as it requires high geo-temporal granularity, high coverage across the population, and appropriate adjustment for seasonal and other external effects. Here, we propose a computational methodology, building on Maslow’s hierarchy of needs, that can capture a holistic view of relative changes in needs following the pandemic through a difference-in-differences approach that corrects for seasonality and volume variations. We apply this approach to characterize changes in human needs across physiological, socioeconomic, and psychological realms in the US, based on more than 35 billion search interactions spanning over 36,000 ZIP codes over a period of 14 months. The analyses reveal that the expression of basic human needs has increased exponentially while higher-level aspirations declined during the pandemic in comparison to the pre-pandemic period. In exploring the timing and variations in statewide policies, we find that the durations of shelter-in-place mandates have influenced social and emotional needs significantly. We demonstrate that potential barriers to addressing critical needs, such as support for unemployment and domestic violence, can be identified through web search interactions. Our approach and results suggest that population-scale monitoring of shifts in human needs can inform policies and recovery efforts for current and anticipated needs.

Web search; Search logs; Human needs; Population-scale; Pandemic; COVID-19; Public policy; Difference-in-differences

^†^†ccs: Information systems Web log analysis^†^†ccs: Human-centered computing

1. Introduction

Many of the existing studies and datasets of the COVID-19 global pandemic focus on the biomedical and epidemiological aspect of the case and fatality rates, including efforts in detection, infection propagation, therapeutic intervention, and vaccine design, with a gaze fixed on the virus and illness that it causes. Despite the direct focus on mitigating the spread and morbidity of infection (Wellenius et al., 2020), pandemic-related policy decisions and investments cannot be made on health information alone. Challenges span societal (e.g., disparities (Chowkwanyun and Reed Jr, 2020)), economic (e.g., unemployment (Baek et al., 2020; Coibion et al., 2020)), and psychosocial (e.g., stress, anxiety (Pfefferbaum and North, 2020), loneliness (Usher et al., 2020)) realms. Recent work has called for identifying and understanding the multi-level system of humans needs and well-being for pandemic response and recovery strategies (Ryan et al., 2020). Our goal is to better understand the influences of the pandemic and associated policy decisions on a multitude of human needs, where new insights about shifting needs can guide valuable refinements of policies and motivate the development of new interventions, programs, and investments.

Quantifying human needs across the population is important but challenging, as it requires innovative, ethical, privacy-preserving approaches with fine-grained and broad geo-temporal coverage. A standard way to assess human needs is through survey-based measures (Gallup, 1976), which can be costly and time-consuming to conduct at large scales. Surveys are difficult to manage across time and geographies when the desire is to provide fine-grained analyses longitudinally and to be able to understand and react in near real-time. Passively observing human behaviors is another approach, grounded in the fact that physical and psychosocial needs motivate human behaviors to express and fulfill those needs when they are unmet (Maslow, 1943). For example, historical purchase behaviors are used to predict future consumer needs in market research (nie, 2020), but this approach is also limited to smaller-scale, consumer and commercial interests. E-commerce platforms (e.g., Amazon marketplace) or specialized service providers (e.g., Talkspace, Coursera) may have access to large-scale, real-time analysis of customer behaviors, but they are focused tightly on specific needs. Publicly available social network data (e.g., Twitter), have been used to characterize needs (Alharthi et al., 2017; Yang and Li, 2013), but these studies examine a subset of needs from data that only portrays externalized behaviors. Such fragmentation of data limits the capture of broader expressions and comparisons across a broad spectrum of human needs.

We address these limitations in obtaining signals about human needs by observing the behaviors of people through their everyday interactions with a web search engine. Human behaviors, through which human needs are expressed or fulfilled, often involve seeking information or obtaining tangible support or material items, for which web search has been an integral component. Thus, search logs provide a unique lens into human needs via providing signals about human behaviors in their natural state, at large scale, and on already routinely collected data.

We propose a computational methodology built on constructs of human needs by Maslow (Maslow, 1943, 1970) and Max-Neef (Max-Neef, 1992). We characterize pandemic changes across a broad spectrum of fundamental human needs, spanning five broad human needs categories–Self Actualization, Cognitive, Love and Belonging, Safety, and Physiological–and 79 subcategories. We apply this framework to a dataset of 35+ billion search interactions across 36,000+ ZIP codes in the United States and over 14 months (7 months in 2019 and 7 months in 2020) to map search query strings and click interactions to human needs, resulting in over three billion expressions of human needs. We demonstrate how this approach enables the examination of shifts in fundamental human needs based on disruptions induced by the pandemic.

Our contributions include the following:

•

We propose a novel computational framework for characterizing a holistic view of human behaviors, intents, and unmet needs based on web search logs and human motivational theories (Sec. 4).
•

We leverage a difference-in-differences approach (Lechner et al., 2011) to quantify the impact of the pandemic and its associated policies on the relative changes in needs while controlling for seasonality and external factors (Sec. 4.4).
•

We present the first population-scale analysis across a holistic set of human needs during the COVID-19 pandemic in the United States through the use of web search logs (35+ billion search interactions for 14 months on 86% of US ZIP codes; Sections 5-8).
•

We find that search interactions in pursuit of basic human needs (i.e., Physiological, Safety) have increased exponentially during the pandemic while several higher-level aspirations (i.e., Self Actualization, Cognitive) have declined (Sections 5, 6).
•

We observe geographical differences in how differing statewide shelter-in-place policies are associated with short-term and long-term changes in social and emotional needs (Sec. 7).
•

We demonstrate that potential barriers to accessing critical resources, in support of people facing unemployment or domestic violence, can be identified through search interactions combined with external data sources (Sec. 8).

Our work suggests that signals from web search logs can be used to characterize and to monitor over time human needs at a population scale. Our findings also emphasize the importance of tracking broad sets of human needs in combination with other reported measures to identify gaps in our current understanding of challenges and support, to measure the impact of policy changes, and to design policies and programs that can meet the needs.

2. Related Work

Quantifying Human Needs. Theories about basic human needs have been discussed for close to a century (Pittman and Zeigler, 2007). In particular, Maslow’s hierarchy of needs (Maslow, 1943, 1970) has been applied in numerous domains (Yang and Li, 2013; van Lenthe et al., 2015; Cerbara et al., 2020; Moss et al., 2010), despite criticisms of the validity of the theory (Wahba and Bridwell, 1976). These theories are aimed at providing a holistic understanding of human needs. The characterizations of broad spectrum of needs from these models are increasingly relevant during the pandemic (Ryan et al., 2020). Most studies use survey-based methods to measure human needs (Cerbara et al., 2020; Tay and Diener, 2011; Mathes and Edwards, 1978; Mitchell and Moudgill, 1976). Others have applied human needs theories in computational social science (Alharthi et al., 2017; Yang and Li, 2013; Long et al., 2020), but they focus deeply on specific topics (e.g., consumer behavior, well-being) and leverage publicly available social network data. In contrast, we introduce a computational methodology to extract a full spectrum of human needs at population scales, which is critical for aligning policies with societal needs that they are intended to support.

Web Search Logs for Human Needs Analysis. In addition, we harness search interaction data which captures more natural observations of human needs (Dumais et al., 2014). Web search logs have been used to understand human behaviors across many different domains (Althoff et al., 2018; Ginsberg et al., 2009; West et al., 2013a; Weber et al., 2012), time (Althoff et al., 2016, 2017; Paul et al., 2016; Fourney et al., 2015), location (West et al., 2013b; Sadilek et al., 2018), and to predict the now and the future (D’Amuri and Marcucci, 2010; Choi and Varian, 2012), but prior studies typically focus on a single aspect of human well-being. Google Trends APIs have helped to stimulate a prolific range of research in the context of the COVID-19 pandemic for physical (Lin et al., 2020), psychological (Tubadji et al., 2020), and socioeconomic (Gupta et al., 2020; Abay et al., 2020) well-being that operate on highly normalized and aggregated data. In distinction, our work leverages fine-grained geospatial comparisons across 79 need subcategories and a difference-in-differences methodology, which allows for controlling confounders and understanding national coverage with flexible geo-temporal normalization and aggregation. We also improve detection of needs by leveraging click interactions.

3. Dataset and Validity

Refer to caption — Figure 1. Illustration of human needs detection framework. (a) Human needs are represented by a ladder according to Maslow’s hierarchy of needs to indicate than a person may have multiple needs simultaneously. (b) These needs are expressed through search interactions, which can be categorized through keyword matches and/or subsequent clicks into relevant search result pages. (c) Each search behavior is then aggregated across categories of human needs, time, or geography. (d) To quantify changes in needs, aggregated needs are compared between pre-pandemic and pandemic periods while adjusting for seasonal and query volume variation.

3.1. Dataset, Privacy, and Ethics

We collected a dataset containing a random sample of deidentified search interactions from the first seven months of the years 2019 and 2020 obtained from Microsoft’s Bing search engine. For each search interaction, we collected the search query strings, all subsequent clicks from the search results page (e.g., clicked URL), and time and ZIP code location of the search interaction. The resulting dataset contains 35+ billion search interactions and represents the web search traffic of over 86% (36,667/42,632) of US ZIP codes associated with at least 100 queries per month so as to preserve anonymity (Table 1). All data was deidentified, aggregated to ZIP code levels or higher, and stored in a way to preserve the privacy of the users and in accordance to Bing’s Privacy Policy. Our study was approved by the Microsoft Research Institutional Review Board (IRB).

Observation period	14 months (Jan 1-Aug 2 in 2019-2020)
# of queries	35,650,687,581
# of human need queries	3,250,228,644
# of days	428
# of ZIP codes	36,667

Table 1. Descriptive statistics for our web search dataset.

3.2. Validation of Data

Considering potential threats to validity, we examined the dataset from three perspectives: coverage of the population, analysis of selection bias, and reliability of the trends.

Analyzing National Representation. To understand how much of the US population is represented by the collected data, we obtained demographics data from the Census Reporter API (Bureau, 2018). The demographics of the ZIP codes in our dataset closely matched the US population demographics (Appendix A.1). Although query volumes are not uniformly distributed across these ZIP codes, the vast majority of the ZIP codes are included in our dataset. We leverage location information for our analysis when geographical differentiation is necessary.

Analyzing Selection Bias. To understand potential biases in socioeconomic circumstances that would influence the usage of the Bing search engine, we leveraged deidentified client id as a proxy for a unique user to estimate the ‘client rate’, or how much of the population in a ZIP code is using the Bing search engine. We examined the correlation between the client rate and various demographic factors. While the factors describe some of the variance, none were correlated more strongly than $r$ =–0.058 (% Housing Owned), suggesting that the dataset is not strongly biased towards any single demographic (Appendix A.2).

Reliability of Search Interaction Trends. Many Americans use other search engines such as Google. Therefore, we compared search trends for Bing with data available via the Google Trends API for the same time period and for specific keywords in each need category. We found that the search trends are highly similar, with a median Pearson correlation of 0.96 (min=0.45, max=0.98, all $p$ ¡0.001). This implies that our findings are not simply an artifact of using one search engine over another (Appendix A.3).

4. Human Needs Framework

4.1. Human Needs Categories

We draw inspiration from Maslow’s hierarchy of fundamental human needs (Maslow, 1943, 1970) to tag each search interaction with one or more of five broad categories of needs. We represent human needs as a ladder to convey that a person may have multiple needs (Fig. 1a). Safety and Physiological are considered as ‘basic’ needs. Love and Belonging is often considered to be ‘psychological’ needs, and Cognitive and Self Actualization are considered as ‘growth’ needs, defined in more detail below:

Self Actualization needs are about realizing personal potential, seeking personal growth, and self-fulfillment. Topics include: hobbies; parenting; wedding; talent acquisition; goals; charity.

Cognitive needs are about pursuing knowledge and intelligence through learning and discovering. Topics include: online education, learning materials; educational degrees; cognition; memory; focus.

Love and Belonging are social and emotional needs and include emotionally-based relationships such as friendship, family, dating, sexual intimacy. Topics include: mental health or emotions; social network or activities; relationships, dating, divorce or breakup.

Safety needs stem from our desire to seek order, stability, and protection from elements in the world. Topics include: personal protection; finances; banking; job search; unemployment; housing.

Physiological needs are the basic animal needs such as air, food, drink, shelter, warmth, sex, and other body needs. Topics include: health; food and groceries; basic staples; sleep; transportation.

To understand the nuances of the needs, we further subdivided the five main categories into 79 subcategories. Several researchers independently developed subcategories which were combined and resolved collaboratively through consensus meetings. Appendix A.4 describes this process in detail with the full taxonomy of our need categories, example queries and/or clicked page URLs.

4.2. Human Needs Detection

A search interaction can be an observation of the underlying human need in two ways: (1) an expression of a potential satisfier (physical or information) for that need, or (2) a direct expression of the deficiency or satisfaction of that need. For example, a search query for ‘bandages’ with a subsequent click on ‘amazon.com’ could indicate a purchase intent that satisfies a Physiological need. We require the additional click into one of many e-commerce domains to solidify that this interaction is a purchase intent. Information search about ‘online games with friends’ could satisfy a Love and Belonging need. A need (satisfaction or deficiency thereof) could be directly expressed in experiential statements such as ‘I feel depressed’ (Love and Belonging).

We match each search interaction to a corresponding need subcategory through simple detectors based on regular expressions and basic propositional logic. Each need subcategory could have multiple regular expressions applied to either the query string, the clicked URL, or both, depending on the complexity of the expression and the need subcategory. We arrived at these regular expressions based on our data through several collaborative consensus meetings until we were satisfied with precision and recall (Appendix A.4). Overall, 9.1% of our query samples matched at least one of the need categories. Each search interaction can satisfy multiple human needs (Max-Neef, 1992), so we allowed each search interaction to be tagged by multiple need categories (only 0.32% have multiple tags). We then aggregated matched search interactions across need categories and subcategories, time (e.g., day, week), and geography (e.g., ZIP code, county, state). Fig. 1 illustrates these steps in detecting and processing of human needs.

Throughout the paper, we report needs as expressed through search interactions and not actual underlying human needs. Some human needs are expressed well by search interactions while other needs are more appropriately expressed through other digital and/or non-digital means. When analyzing changes in such search interactions, careful consideration is required to differentiate actual shifts in the underlying human needs (e.g., health-related needs have increased due to COVID-19) from the equally meaningful changes observed in logs based upon shifts from offline to online behaviors (e.g., online grocery purchases due to store closures), and we present our results in both possible contexts.

4.3. Framework Validation

Our goal is to detect with high precision a large number of needs across a broad set of categories.

Precision. We sampled 1.2 million search interactions that matched at least one need category as a candidate set. From this sample, we randomly chose 100 unique tuples of search query string and clicked URL (e.g., ‘15 lb dumbbells’ and click on ‘walmart.com’) for each of the five high-level need categories, for a total of 500 unique tuples representing 1,530 search interactions in our evaluation set. We selected unique tuples to avoid duplication in labeling, but we mapped the labels back to the original 1,530 search interactions to compute precision on the distribution of the source evaluation set. We then collected human labels for each tuple via Amazon’s Mechanical Turk, where each tuple could be tagged with none, one, or more of the five needs categories. All labels and predictions have Boolean values with no ranking among needs categories.

Upon inspecting the label quality, we found common systematic label errors such as labeling ‘recipe’-related queries as Cognitive needs, or ‘divorce’-related queries or visits to specific government unemployment sites as Physiological needs, where the workers mislabeled the queries according to the definitions we specified in the task detail. Other errors were due to inherent ambiguities in search. For example, ‘rent in florida coronavirus’ is tagged as Physiological for ‘coronavirus’ but not as Safety for ‘rent’ because our high precision detector requires more qualified keywords such as ‘apartment rent.’ Although the worker tagged this as Safety (i.e., rent for shelter), the use of the word ‘rent’ here may not be shelter-related. We took a conservative approach of only correcting definitive label errors and not ambiguous errors, and our evaluation set achieved a precision of $97.2\%$ , using the example-based precision metric defined in (Zhang and Zhou, 2013) for multi-label classification.

Recall. Although it is infeasible to ensure a perfect recall across a massive dataset, it is important that we capture a significant number of needs expressions. We find that 9.1% of our search interactions match at least one need category. While this recall is significant and led to more than 3.2 billion detections of needs expressions, we note that high recall is not necessary for an unbiased analysis approach, because we conduct a fair comparison among the outputs of the same detectors across pre-pandemic and pandemic periods. We did not find that the exact expressions of needs varied drastically within our dataset that would indicate any temporal bias. We also investigated whether our need expressions were dominated by a few categories. Clicks to YouTube or Facebook dominated, but still only represented 1% of our dataset. We categorized visits to these sites based on their primary functions (i.e., Facebook for social networking and YouTube for media consumption). We found that our results were robust, whether or not we included these high-traffic sites in our dataset.

4.4. Quantifying Changes in Human Needs

Our goal is to quantify the change in human needs during the pandemic relative to the pre-pandemic period. This can be challenging due to potential confounding effects of yearly seasonal variations, weekly seasonal variations, and variations in query volume over time. Conceptually, we control for yearly seasonal effects through comparisons with the previous year, for weekly seasonal effects by matching the day of the week between both years (i.e., Mon Jan 6, 2020 is aligned to Mon Jan 7, 2019), and by considering relative proportions of the query volume represented by each need over time. Formally, we follow a difference-in-differences methodology (Lechner et al., 2011; Dimick and Ryan, 2014), commonly used in economics, to account for confounding effects between comparison groups. Finally, our adjusted effect size is the logarithm of the ratio between two groups. This is effectively the difference-in-differences approach applied to the logarithmic effect sizes and has the advantage of the effect sizes having symmetric properties (i.e., $\Delta(t_{1};t_{2})$ = $-\Delta(t_{2};t_{1})$ ) (Graff, 2014; Cole and Altman, 2017). This step allows for appropriate comparison of effect sizes across both increases and decreases in need. Our estimate of the relative change in human need $C$ between two time periods is defined as

C(t_{1};t_{2},n)=\log_{2}\left(\frac{E(t_{2}^{2020},n)}{E(t_{1}^{2020},n)}\right)-\log_{2}\left(\frac{E(t_{2}^{2019},n)}{E(t_{1}^{2019},n)}\right)

where $E(t_{2}^{2020},n)$ is the expression of need $n$ at some time $t_{2}^{2020}$ in 2020 (i.e., after the pandemic declaration) and $E(t_{1}^{2020},n)$ is the expression of need at $t_{1}^{2020}$ (i.e., before the pandemic).

Across all following analyses, we choose the mean daily expression of needs between Jan 6 to Feb 23, 2020 as the ‘pre-pandemic baseline’, referred to throughout the paper, and dates on or after Mar 16 as the ‘pandemic period’ because individual states declared a state of emergency at different times (Feb 29 to Mar 15). We then compute the 95% confidence interval on this multiplicative effect size by using bootstrap resampling with replacement (N=500). We report mean estimates and p-values throughout the text and 95% confidence intervals in all figures and tables where applicable. All time series figures (Figures 2, 3, 5B) indicate the moving average of the daily relative changes, computed from 3 days before to 3 days after.

5. Temporal Changes in Human Needs

We first consider how human needs change over time across the US in the context of major events surrounding the pandemic. We compute the daily relative change in the expressed needs in comparison to the pre-pandemic baseline, as described in Sec. 4.4, for the duration of our entire dataset, giving us per-day relative changes in all need categories and subcategories. For each inflection point and major national event, we examine need subcategories with the highest relative changes to understand which contribute the most to the overall need.

Elevated Needs and Contributing Subcategories. Fig. 2 illustrates daily relative changes of needs on a log scale, where zero indicates no change. Overall, we see that all need categories were at elevated rates during March through May relative to the earlier months. A few of the local inflection points correspond to US national events, such as the declaration of national emergency on Mar 13 or the first stimulus checks being deposited on Apr 11.

Physiological needs start to increase first around February (Fig. 2A), dominated by health condition related queries ( $C$ = $1.46$ on Feb 29) and subsequently by toilet paper purchase and health measurement equipment purchase ( $C$ = $1.14$ , $0.76$ on Mar 6 respectively; Fig. 2B). Around Mar 16, Physiological needs peak at over 3.8 times the baseline ( $2^{1.91}$ ; Fig. 2C). Following national emergency declaration (Mar 13) and mandated lock downs (first on Mar 21), we see a sharp increase in Cognitive needs (Fig. 2D), dominated by educational site visits and online education queries ( $C$ = $1.97$ , $1.42$ on Mar 23). Self Actualization needs peak around Apr 11 (Fig. 2E), dominated by cooking site visits and cooking related queries ( $C$ = $1.73$ , $1.15$ ), and online social activities queries and social technology uses dominate Love and Belonging needs ( $C$ =1.77, 1.72 on Apr 11). A sharp spike of Safety needs can be seen shortly after the first stimulus checks were deposited: stimulus related queries, state unemployment site visits, COVID-19 protection purchase dominate Safety needs ( $C$ =8.17, 5.49, 5.12 on Apr 18; Fig. 2F). While other needs start to trend downwards or stabilize throughout much of May-July, Physiological needs increase for a second time with additional interests in health conditions, followed by Safety needs with queries related to economic stimulus and loans (Fig. 2G), which aligns with the rise of COVID-19 cases in the US around Jun 6.

Shifting of Needs. Based on the severe health impacts of the COVID-19 pandemic, we expected to see and confirmed that Physiological needs dominate throughout our dataset, as COVID-19 is still a major US public health issue at the time of writing. At a glance, we see two instances of the surge in Physiological needs followed by a subsequent increase in Safety needs. As Physiological concerns rise, public health responses (e.g., business closures or restrictions) could potentially induce instabilities in Safety needs, and this observation needs to be further investigated. We see basic needs expressed before other needs consistent with the hypothesis by Maslow (Maslow, 1943) and observations by others (Tay and Diener, 2011). We also expected to see a decrease in the expression of growth needs (Self Actualization and Cognitive) as people’s attention shifts toward basic needs. However, both Cognitive and Self Actualization needs increased overall, with the increase in Cognitive needs being more temporary and Self Actualization being more sustained. Despite health and economic concerns, interests in recreational activities or hobbies (e.g., cooking, gaming) contribute to this steady 23% (= $2^{0.3}$ –1) increase in Self Actualization needs. Further research into the temporary nature of Cognitive needs and the long-term impact of such sustained interest in Self Actualization is necessary.

We see that the peak in Physiological needs occurs around four weeks before the peak in Safety needs (Fig. 2C, F), while the second set of peaks are a few days apart (Fig. 2H). This could be an indication of phenomena like resilience or endurance from economics and disaster management that requires further examination (Martin and Sunley, 2015; Fletcher and Sarkar, 2013).

6. Significant Changes in Human Needs

Next, we examine individual need subcategories that present the largest increase or decrease in search expressions, possibly due to the pandemic impact. To explore these two ends of the spectrum, we compute the mean relative changes in needs during the initial four weeks of the pandemic period (Mar 16 to Apr 12) compared to the pre-pandemic baseline, as described in Sec. 4.4. We then examined the top 12 need subcategories with the largest increase or decrease in the relative change.

Need	Need Subcategory	$C_{\text{mean}}$	$C_{\text{max}}$	$2^{C_{\text{max}}}$ - $1$
Phys	Toilet paper purchase	6.11 $\pm$ 0.08	7.00	12691.1%
Safe	Stimulus related queries	5.69 $\pm$ 0.03	8.17	28601.9%
Safe	Unemploy. related queries	4.88 $\pm$ 0.02	5.72	5156.4%
Safe	State unemploy. site visits	4.81 $\pm$ 0.04	6.26	7585.1%
Safe	COVID-19 prot. purchase	4.67 $\pm$ 0.03	5.57	4634.0%
Phys	Health meas. equip. purchase	3.43 $\pm$ 0.04	4.56	2257.2%
Phys	Health cond. related queries	3.12 $\pm$ 0.01	3.54	1065.6%
Phys	Food assist. related queries	2.62 $\pm$ 0.03	3.12	771.8%
Phys	Grocery related queries	2.05 $\pm$ 0.01	2.54	480.8%
Phys	Food delivery queries	1.84 $\pm$ 0.02	2.26	379.7%
L&B	Online social activity queries	1.77 $\pm$ 0.09	2.98	688.4%
Phys	Food delivery site visits	1.7 $\pm$ 0.01	2.26	379.3%

Table 2. Top 12 need subcategories with the largest increase in mean relative change in need within the initial 4 weeks of the pandemic with 95% confidence intervals, maximum relative change in the dataset, and maximum percent change.

Need	Need Subcategory	$C_{\text{mean}}$	$C_{\text{min}}$	$2^{C_{\text{min}}}$ - $1$
SA	Wedding related purchase	-1.49 $\pm$ 0.03	-1.76	-70.4%
SA	Wedding site visits	-1.25 $\pm$ 0.02	-1.56	-66.2%
Cog	Edu. degree related queries	-0.87 $\pm$ 0.06	-1.09	-53.1%
Safe	Housing related queries	-0.71 $\pm$ 0.05	-1.09	-53.2%
Safe	Job search related queries	-0.65 $\pm$ 0.03	-0.91	-46.7%
Safe	Job search site visits	-0.61 $\pm$ 0.02	-0.95	-48.1%
Phys	Apparel purchase	-0.60 $\pm$ 0.01	-0.84	-44.1%
SA	Outdoor related queries	-0.59 $\pm$ 0.01	-1.07	-52.3%
SA	Life goal related queries	-0.57 $\pm$ 0.09	-1.23	-57.2%
Safe	Domestic violence queries	-0.54 $\pm$ 0.04	-0.97	-49.0%
Safe	Rental related queries	-0.53 $\pm$ 0.06	-0.82	-43.5%
L&B	Divorce related queries	-0.49 $\pm$ 0.03	-0.93	-47.3%

Table 3. Top 12 need subcategories with the largest decrease in mean relative change in need within the initial 4 weeks of the pandemic with 95% confidence intervals, minimum relative change in the dataset, and minimum percent change.

Heightened Physiological and Safety Needs. Table 2 shows that 11 need subcategories with the most increase fall under Physiological and Safety needs as seen from the temporal trends in Sec. 5, and one (online social activities queries) belongs to the Love and Belonging need category. Toilet paper purchase reached a maximum increase of 127 times the pre-pandemic baseline (Mar 16). Recall that these are not just queries containing ‘toilet paper’, but purchase intents as indicated by subsequent clicks to e-commerce sites (Sec. 4.2). Such a high level of interest in toilet papers is commonly attributed to panic buying due to the supply scarcity (Hamilton, 2020) and media coverage (Garfin et al., 2020). Stimulus related queries, including general terms like ‘loan forgiveness,’ reached an even higher maximum increase of 286 times the baseline (Apr 18) and is sustained at that high level through July, reflecting the magnitude of the pandemic’s impact on the US economy.

When we examine the daily trends, we see that COVID-19 protection purchase exhibits a small peak ( $C$ =1.0 on Jan 29) after the first reported COVID-19 case (Jan 20), and the needs quickly escalate from Feb 20 (Fig. 3B), at least three weeks earlier than other needs (Fig. 3A,C,D) that do not escalate until the national emergency declaration. A subsequent peak on Apr 3 ( $C$ =5.6) coincides with CDC’s updated recommendation on cloth-based mask use. We also find that indicators of social-economic instabilities such as unemployment site visits and food assistance related queries still have not returned to their baseline levels (Fig. 3A,C), arguably because the pandemic is still in effect. Online social activities queries follow a similar pattern (Fig. 3D), reflecting the need to satisfy lock-down induced social isolation through online services, but it also raises the question of potential permanent shifts in ways of satisfying social needs. Our findings revealed that only a few of these needs have returned to the pre-pandemic baselines while many of them are sustained at elevated rates.

Shifts Away from Positive Outlooks.

Table 3 shows the most decrease in the expression of several Self Actualization and one Cognitive need subcategories. Specifically, indications of Self Actualization needs for partnership have declined by more than 64% (= $2^{-1.49}$ –1) of their baseline throughout the typical US wedding season around Spring and early Summer, which is expected given restrictions on large gatherings. In addition, needs that are typically associated with growth, positive outlook, or new opportunities have taken a large toll. In Self Actualization, queries about life goals also see a large decline. In other need categories, needs expressed by educational degree related queries, job search related queries, job search site visits, or housing related queries have declined by over 34% (= $2^{-0.61}$ –1) of their baseline.

Upon inspection of daily trends, expressions of forward-looking needs have decreased and remain below the pre-pandemic baseline (Fig. 3E,F,G). The sustained decline in job search related queries (Fig. 3G) juxtaposed with near 30 times increase in unemployment needs (Fig. 3A) is a troubling evidence of the declining labor force as seen in other studies (Coibion et al., 2020). Indications of growth interests in educational degrees or life goals have not recovered (Fig. 3F). These results combined with heightened Physiological needs suggest a shift of focus away from individual growth. Divorce related queries exhibited a maximum of 47% (= $2^{-0.93}$ ; Fig. 3H) decline, possibly reflecting the challenges that families face in proceeding with divorce during the lock downs (Lebow, 2020). Therefore, underlying mechanisms for these shifts, long-term impact of the lack of growth needs, and support for relationships should be further studied. See Appendix A.5 for corresponding figures on all 24 subcategories.

7. Geographical Differences in Needs

We shift our attention to how the pandemic and its related policies differentially influence local subpopulations. We use a set of statewide policies¹¹1Although our dataset allows ZIP code level analysis, a comprehensive list of local policies across the US are difficult to obtain at ZIP code, city, or even county levels. readily available through the COVID-19 US State Policy Database (Raifman et al., 2020). We examine the impact of the shelter-in-place mandate (its duration and effective date) on social-emotional and relationship needs (i.e., eight Love and Belonging and two wedding-related Self Actualization subcategories). We hypothesized that social isolation induced by longer shelter-in-place mandates generates more expressions of social-emotional needs.

We used the date on which the shelter-in-place mandate was enacted and relaxed or lifted for 38 states to derive the mandate duration²²2Only 38 states had the start and end dates in the dataset as of this writing.. We compute two relative changes in needs for each state. First, to understand the short-term impact of the shelter-in-place mandate, we compute the relative change in needs for two weeks after each state’s shelter-in-place mandate compared to the one week before the mandates³³3We chose one week before the mandate because it is the maximum number of full weeks after the declaration of national emergency and before the earliest start date (Mar 21) for any shelter-in-place mandate.. Second, to understand the long-term impact of the mandate, we compute the relative change in needs between the pre-pandemic baseline and the last four weeks of our dataset (Jul 6 to Aug 2), which is at least two weeks after the last state lifted its mandate (Jun 19). To quantify the potential impact of these mandates, we ran a Pearson correlation analysis (1) between the start date of the shelter-in-place mandates (i.e., ISO day number) and the short-term relative changes in needs, and (2) between the duration of the mandates and the long-term relative changes in needs.

Early Adjustments to Social Needs. The start date of shelter-in-place mandates ranged between Mar 21 and Apr 7. We find that the relative changes in online social activities queries ( $r$ = $-0.53$ , $p$ ¡0.001), wedding site visits (( $r$ =0.48, $p$ =0.002), wedding related purchase ( $r$ =0.43, $p$ =0.006), and mental health resource site visits ( $r$ =0.47, $p$ =0.003) needs were most correlated with the start date. People from states that have earlier shelter-in-place mandates expressed significantly reduced interests in weddings and mental health site visits (Fig. 4A) and significantly more need for online social activities (Fig. 4B). For example, in the two weeks after the mandate, people in New Jersey (mandate on Mar 21) sought online mental health resources 25.7% less than the week before the mandate, while people in South Carolina (mandate on Apr 7) sought those resources 13.4% more than the week before.

Long-term Social Impact. When we examine the long-term changes in expressed needs, we find that shifts in negative mental health experiential queries ( $r$ =0.42, $p$ = $0.010$ , Fig. 4C), wedding site visits ( $r$ = $-0.37$ , $p$ =0.022, Fig. 4D), and wedding related purchase ( $r$ = $-0.35$ , $p$ =0.033) were most correlated with the duration of sheltering and closures. The duration of the shelter-in-place mandates were highly correlated with the start date ( $r$ = $-0.62$ , $p$ ¡0.001), indicating that people faced with earlier mandates are also impacted by longer mandates. As we discovered above, these people likely suppressed their needs for weddings or mental health support early during the pandemic. At the same time, as wedding needs slowly recover to the pre-pandemic baseline (Fig. 3E), those impacted by longer mandates are even slower in their recovery of wedding needs and are more likely to express negative mental health issues. For example, people in Mississippi (24 days of shelter-in-place) expressed 33.2% less negative mental health experiences than before the pandemic while people in Oregon (88 days) expressed 27.2% more negative mental health experiences than before (Fig. 4C). Others attribute this increase in negative emotions during the pandemic to the shift to basic needs (Cerbara et al., 2020). Per-state analysis of this shift along with differential prevalence of COVID-19 may be necessary to understand the mechanisms for this increase in negative emotions and to provide appropriate social-emotional support.

8. Gaps between Expressed and Reported Needs

As we have demonstrated so far, there are many needs that are well expressed by web search interactions such as purchasing goods online, looking for health information online, or accessing economic assistance through government websites. Our final analysis examines a gap between how web search facilitates the expression of these needs and the reported fulfillment of these needs to discover potential barriers to accessing critical resources. To demonstrate how our approach allows for exploring these barriers, we focus on two need subcategories: unemployment and domestic violence.

Expressed Unemployment Needs vs. Reported Claims. We obtained weekly, seasonally adjusted initial unemployment claims for 2020 from the US Department of Labor⁴⁴4https://oui.doleta.gov/unemploy/claims.asp and computed the relative changes in unemployment claims using the same approach as in Sec. 4.4. We compare the change in reported unemployment claims with the change in expression of unemployment needs in search and find that these two changes are closely aligned (Pearson $r$ =0.996, $p$ ¡0.001) (Fig. 5A). Despite policies that extend unemployment eligibility for up to 26 weeks during the pandemic, we see that the expression of unemployment needs remain at 25% higher than the reported claims since April. This discrepancy corroborates with known issues with the unemployment benefits: many people were confused about the benefits (e.g., job search requirement) and were being denied, having to search for more information or file multiple applications (Schwartz et al., 2020; Bhardwaj, 2020). Although our analysis excludes those who use traditional methods for filing unemployment claims (e.g., mail, phone), our results indicate that web search is a critical resource that facilitates unemployment needs and highlight a potential gap in satisfying unemployment needs which requires further investigation.

Expressed Domestic Violence Needs vs. Reports. One of the dire consequences of the COVID-19 pandemic is an increased risk of domestic violence due to shelter-in-place mandates exacerbated by physical, financial, and social-emotional stressors and increased alcohol consumption at home (Campbell, 2020; Li and Schwartzapfel, 2020). Our analysis shows that the expression of needs for domestic violence queries has dropped by nearly 36.7% (= $2^{-0.66}$ –1) since the pandemic and stabilized at around -15.9% (= $2^{-0.25}$ –1, Fig. 5B) below pre-pandemic levels. Interests in firearm purchase have increased by over 40% (= $2^{0.5}$ –1), corroborating other reports of increased gun sales and gun violence and worrisome links to fatal domestic violence incidents (Sutherland et al., 2020; Hatchimonji et al., 2020; Campbell, 2020).

A similar decrease in domestic violence related metrics in March is seen by national hotlines⁵⁵5https://crisistrends.org/; https://www.thehotline.org/. Some reports point to potential underreporting by victims unable to reach out for help under constant surveillance by the offenders at home and fearful of exposure to the virus at a shelter (Bullinger et al., 2020; Southall, 2020). Although the expansion of online resources are critical to addressing the rise in domestic violence (Mahase, 2020), many studies highlight the need to recognize control tactics that prevent access to these digital resources (Ross and Aitken, 2020). The sustained decrease in domestic violence queries seen in our data may indicate such barriers or shifts in access methods for these online resources. The complexities of domestic violence are evident in the inconsistencies across many data sources, resource media, and contexts. Our results frame an urgent question that needs to be resolved through expertise in domestic violence and social work organizations and through careful combination of multiple data sources to identify the underlying explanations for our findings.

9. Discussion and Conclusion

We presented a computational methodology based on theories of human needs to quantify the effects of the pandemic and its related events on web search interactions. Although all need categories had elevated changes in needs, we found that basic need subcategories were elevated the most while growth-based need subcategories, indicative of positive outlooks in life, were subdued (Sections 5, 6). We also found that earlier and longer shelter-in-place mandates may come with an unintended impact on mental health needs (Sec. 7). We used unemployment and domestic violence related queries to demonstrate how our methodology could help expose gaps between the expression and fulfillment of needs–and frame questions and directions for urgent investigation (Sec. 8).

Limitations. We cannot use our method to make causal claims. If other major, concurrent events (e.g., Black Lives Matter) have significantly influenced human needs during the observations window, we are not be able to distinguish between needs caused by the pandemic or these events. However, our analysis clearly presents changes in the expression of human needs which can be important on its own for understanding rising needs and in considering factors for the design of specific interventions. Our analysis relies on heuristics and correct inference of needs from search interactions. But our framework is theory-driven, comprehensive, and achieves 97% precision across five main need categories. Further fine-tuning of the framework should be motivated by precision-recall requirements of individual analysis goals. Although expressions of needs may not reflect the actual underlying human needs, the methodology serves as a useful detector for needs, as we have demonstrated through its corroboration with several external findings, and combining our methodology with additional data sources and interdisciplinary collaboration can further the understanding of changes in human needs during global crises. In contrast to traditional methods of measuring needs, our approach can operate at population scales relying solely on data that is already routinely collected by web search engines, and thus allows for effective, retrospective, and fine-grained studies with observation periods of over one year.

Implications: Resilience and Vulnerability. Our work has demonstrated major shifts in a spectrum of human needs during the pandemic, echoing the need to understand the societal, economic, and psychosocial effects of these events and the underlying system of human needs. Such understanding is essential for guiding programs around outreach and support of people during the current pandemic and preparing for future pandemic responses and recovery efforts (Ryan et al., 2020). We observed that some needs have returned to their baseline levels while others are sustained at elevated (e.g., health and economic instability) or suppressed (e.g., job search, educational degree) levels (Fig. 3). As we saw in Sec. 5, temporal variations on how and when certain needs arise in response to major disruptions could indicate a degree of psychological and economic resilience (Martin and Sunley, 2015; Fletcher and Sarkar, 2013). Understanding how well a region can endure disruptions is crucial for balancing health risks against societal risks, especially for vulnerable populations who are severely impacted by the pandemic (Chowkwanyun and Reed Jr, 2020). Our framework could be used to quantify how much a community has been and might be able to endure social and economic distress. For example, a region may be able to withstand prolonged business closures before loan forgiveness needs surge, while another region may show immediate need for financial stimulus. In addition, when coupled with demographic variables, our framework could be used to examine disparities in needs which may help with understandings of the disparate impacts of blanket policies on the most vulnerable populations.

Implications: Preparedness and Resources. Our work revealed strong correlations indicative of a long-term mental health impact of longer shelter-in-place mandate, suggesting that states with longer duration of shelter-in-place mandates may need to provide additional support for social emotional well-being (Fig. 4). We also saw potential barriers to accessing critical resources for unemployment and domestic violence; a decrease in domestic violence queries in spite of conflicting anecdotes should raise questions and an alarm (Fig. 5). These examples suggest that expressions of needs through web interactions could highlight resource deficiencies or barriers. As regions prepare recovery efforts from COVID-19 or make plans for future disruptions, our methodology can be harnessed to monitor or anticipate a spectrum of needs at various geotemporal granularities; this could help reveal impacts of the pandemic (e.g., mental health (Czeisler et al., 2020)) and cater the appropriate interventions that meet individual and societal needs.

Future Directions. We see this work as a step towards achieving a more holistic understanding of the multiple influences of the pandemic and associated policies and programs. While we focused on retrospective analyses, our approach has the potential to be used in near real-time to monitor human needs and support future policy decisions. Understanding the disparate impacts of the pandemic and its policies on a full spectrum of human needs, especially for vulnerable populations, is critical for designing response and recovery efforts for major disruptions. Our work highlighted future research opportunities and calls for action and collaboration across epidemiology, economy, social sciences, risk management, and more. We look forward to further refinement of the methods presented and hope the work will encourage discussions on observing and addressing changes across a broad spectrum of human needs during a global crisis.

Acknowledgements.

We thank the reviewers and colleagues for their valuable feedback. This research was supported in part by NSF grant IIS-1901386, Bill & Melinda Gates Foundation (INV-004841), and the Allen Institute for Artificial Intelligence.

References

(1)
nie (2020) 2020. A Look at How Home Care Product Claim Preferences Have Shifted Amid the COVID-19 Pandemic. https://www.nielsen.com/us/en/insights/article/2020/a-look-at-how-home-care-product-claim-preferences-have-shifted-amid-the-covid-19-pandemic/
Abay et al. (2020) K. A. Abay, K. Tafere, and A. Woldemichael. 2020. Winners and Losers from COVID-19: Global Evidence from Google Search. World Bank Policy Research Working Paper (2020).
Albert and Tullis (2013) W. Albert and T. Tullis. 2013. Measuring the user experience: collecting, analyzing, and presenting usability metrics. Newnes.
Alharthi et al. (2017) R. Alharthi, B. Guthier, C. Guertin, and A. El Saddik. 2017. A dataset for psychological human needs detection from social networks. IEEE Access 5 (2017).
Althoff et al. (2018) T. Althoff, E. Horvitz, and R. W. White. 2018. Psychomotor function measured via online activity predicts motor vehicle fatality risk. NPJ Digit. Med. 1, 1 (2018).
Althoff et al. (2017) T. Althoff, E. Horvitz, R. W. White, and J. Zeitzer. 2017. Harnessing the web for population-scale physiological sensing: A case study of sleep and performance. In WWW.
Althoff et al. (2016) T. Althoff, T. W. White, and E. Horvitz. 2016. Influence of Pokémon Go on physical activity: Study and implications. JMIR 18, 12 (2016).
Baek et al. (2020) C Baek, P. B. McCrory, T. Messer, and P. Mui. 2020. Unemployment effects of stay-at-home orders: Evidence from high frequency claims data. Institute for Research on Labor and Employment Working Paper (2020).
Bhardwaj (2020) P. Bhardwaj. 2020. Most Job Search Requirements for Unemployment Benefits Are Waived. So Why Do State Websites Say Otherwise? https://money.com/unemployment-benefits-job-search-requirements-coronavirus/
Bullinger et al. (2020) Lindsey Rose Bullinger, Jillian B Carr, and Analisa Packham. 2020. COVID-19 and Crime: Effects of Stay-at-Home Orders on Domestic Violence. Technical Report. National Bureau of Economic Research.
Bureau (2018) U.S. Census Bureau. 2014-2018. American Community Survey 5-year estimates. Retrieved June 8, 2020 from https://censusreporter.org
Campbell (2020) A. M. Campbell. 2020. An increasing risk of family violence during the Covid-19 pandemic: Strengthening community collaborations to save lives. Forensic Science International: Reports (2020).
Cerbara et al. (2020) L. Cerbara, G. Ciancimino, M. Crescimbene, F. La Longa, M. R. Parsi, A. Tintori, and R. Palomba. 2020. A nation-wide survey on emotional and psychological impacts of COVID-19 social distancing. Eur Rev Med Pharmacol Sci 24, 12 (2020).
Choi and Varian (2012) H. Choi and H. Varian. 2012. Predicting the present with Google Trends. Economic record 88 (2012).
Chowkwanyun and Reed Jr (2020) M. Chowkwanyun and A. L. Reed Jr. 2020. Racial health disparities and Covid-19—caution and context. New England Journal of Medicine (2020).
Coibion et al. (2020) O. Coibion, Y. Gorodnichenko, and M. Weber. 2020. Labor markets during the covid-19 crisis: A preliminary view. Technical Report. NBER.
Cole and Altman (2017) T. J. Cole and D. G. Altman. 2017. Statistics notes: percentage differences, symmetry, and natural logarithms. Bmj 358 (2017).
Czeisler et al. (2020) M. É. Czeisler, R. I. Lane, E. Petrosky, et al. 2020. Mental Health, Substance Use, and Suicidal Ideation During the COVID-19 Pandemic – United States, June 2–30, 2020. MMWR Morb Mortal Wkly Rep (2020).
Dimick and Ryan (2014) J. B. Dimick and A. M. Ryan. 2014. Methods for evaluating changes in health care policy: the difference-in-differences approach. Jama 312, 22 (2014).
Dumais et al. (2014) S. Dumais, R. Jeffries, D. M. Russell, D. Tang, and J. Teevan. 2014. Understanding user behavior through log data and analysis. In Ways of Knowing in HCI. Springer.
D’Amuri and Marcucci (2010) F. D’Amuri and J. Marcucci. 2010. ’Google it!’Forecasting the US unemployment rate with a Google job search index. (2010).
Fletcher and Sarkar (2013) D. Fletcher and M. Sarkar. 2013. Psychological resilience: A review and critique of definitions, concepts, and theory. European psychologist 18, 1 (2013).
Fourney et al. (2015) A. Fourney, R. W. White, and E. Horvitz. 2015. Exploring time-dependent concerns about pregnancy and childbirth from search logs. In ACM CHI.
Gallup (1976) G. H. Gallup. 1976. Human needs and satisfactions: A global survey. Public opinion quarterly 40, 4 (1976).
Garfin et al. (2020) D. R. Garfin, R. C. Silver, and E. A. Holman. 2020. The novel coronavirus (COVID-2019) outbreak: Amplification of public health consequences by media exposure. Health Psychology (2020).
Ginsberg et al. (2009) J. Ginsberg, M. H. Mohebbi, R. S. Patel, L. Brammer, M. S. Smolinski, and L. Brilliant. 2009. Detecting influenza epidemics using search engine query data. Nature 457, 7232 (2009).
Graff (2014) C. Graff. 2014. Expressing relative differences (in percent) by the difference of natural logarithms. Journal of Mathematical Psychology 60 (2014).
Gupta et al. (2020) S. Gupta, L. Montenovo, T. D. Nguyen, F. L. Rojas, I. M. Schmutte, K. I. Simon, B. A. Weinberg, and C. Wing. 2020. Effects of social distancing policy on labor market outcomes. Technical Report. NBER.
Hamilton (2020) R. Hamilton. 2020. Scarcity and Coronavirus. J. Public Policy Mark. (2020).
Hatchimonji et al. (2020) J. S. Hatchimonji, R. A. Swendiman, M. J. Seamon, and M. L. Nance. 2020. Trauma does not quarantine: Violence during the Covid-19 pandemic. Ann. Surg. (2020).
Lebow (2020) J. L. Lebow. 2020. The Challenges of COVID-19 for Divorcing and Post-divorce Families. Family Process (2020).
Lechner et al. (2011) M. Lechner et al. 2011. The estimation of causal effects by difference-in-difference methods. Foundations and Trends R in Econometrics.
Li and Schwartzapfel (2020) W. Li and B. Schwartzapfel. 2020. Is Domestic Violence Rising During the Coronavirus Shutdown? Here’s What the Data Shows. https://www.themarshallproject.org/2020/04/22/is-domestic-violence-rising-during-the-coronavirus-shutdown-here-s-what-the-data-shows/
Lin et al. (2020) Y-H Lin, C-H Liu, and Y-C Chiu. 2020. Google searches for the keywords of “wash hands” predict the speed of national spread of COVID-19 outbreak among 21 countries. Brain, Behavior, and Immunity (2020).
Long et al. (2020) Z. Long, R. Alharthi, and A. El Saddik. 2020. NeedFull–A Tweet Analysis Platform to study Human Needs during the COVID-19 pandemic in New York State. IEEE Access (2020).
Mahase (2020) Elisabeth Mahase. 2020. Covid-19: EU states report 60% rise in emergency calls about domestic violence. BMJ: British Medical Journal (Online) 369 (2020).
Martin and Sunley (2015) R. Martin and P. Sunley. 2015. On the notion of regional economic resilience: conceptualization and explanation. Journal of Economic Geography 15, 1 (2015).
Maslow (1943) A. H. Maslow. 1943. A Theory of Human Motivation. Psychol. Rev. (1943).
Maslow (1970) A. H. Maslow. 1970. New introduction: Religions, values, and peak-experiences. Journal of Transpersonal Psychology 2, 2 (1970).
Mathes and Edwards (1978) E. W. Mathes and L. L. Edwards. 1978. An empirical test of Maslow’s theory of motivation. Journal of Humanistic Psychology 18, 1 (1978).
Max-Neef (1992) M. A. Max-Neef. 1992. Human scale development: conception, application and further reflections. Number 04; HC125, M3.
Mitchell and Moudgill (1976) V. F. Mitchell and P. Moudgill. 1976. Measurement of Maslow’s need hierarchy. Organ. Behav. Hum. Decis. Process 16, 2 (1976).
Moss et al. (2010) S. Moss et al. 2010. An introduction to the entertainment industry. The entertainment industry: An introduction (2010).
Paul et al. (2016) M. J. Paul, R. W. White, and E. Horvitz. 2016. Search and breast cancer: On episodic shifts of attention over life histories of an illness. TWEB 10, 2 (2016).
Pfefferbaum and North (2020) B. Pfefferbaum and C. S. North. 2020. Mental health and the Covid-19 pandemic. New England Journal of Medicine (2020).
Pittman and Zeigler (2007) T. S. Pittman and K. R. Zeigler. 2007. Basic human needs. (2007).
Raifman et al. (2020) J. Raifman, K. Nocka, D. Jones, J. Bor, S. Lipson, J. Jay, M. Cole, N. Krawczyk, P. Chan, S. Galea, et al. 2020. COVID-19 US state policy database. (2020).
Ross and Aitken (2020) Stuart Ross and Sophie Aitken. 2020. “If It Hadn’t Been Online I Don’t Think I Would Have Applied”: Applicant Experiences of an Online Family Violence Intervention Order Process. Journal of interpersonal violence (2020).
Ryan et al. (2020) B. J. Ryan, D. Coppola, D. V. Canyon, M. Brickhouse, and R. Swienton. 2020. COVID-19 Community Stabilization and Sustainability Framework: An Integration of the Maslow Hierarchy of Needs and Social Determinants of Health. Disaster medicine and public health preparedness (2020).
Sadilek et al. (2018) A. Sadilek, S. Caty, L. DiPrete, R. Mansour, T. Schenk, M. Bergtholdt, A. Jha, P. Ramaswami, and E. Gabrilovich. 2018. Machine-learned epidemiology: real-time detection of foodborne illness at scale. NPJ Digit. Med. 1, 1 (2018).
Schwartz et al. (2020) N. D. Schwartz, T. Hsu, and P. Cohen. 2020. Stymied in Seeking Benefits, Millions of Unemployed Go Uncounted. https://www.nytimes.com/2020/04/30/business/economy/coronavirus-unemployment-claims.html
Southall (2020) A. Southall. 2020. Why a Drop in Domestic Violence Reports Might Not Be a Good Sign. https://www.nytimes.com/2020/04/17/nyregion/new-york-city-domestic-violence-coronavirus.html
Sutherland et al. (2020) M. Sutherland, M. McKenney, and A. Elkbuli. 2020. Gun violence during COVID-19 pandemic: Paradoxical trends in New York City, Chicago, Los Angeles and Baltimore. The American Journal of Emergency Medicine (2020).
Tay and Diener (2011) L. Tay and E. Diener. 2011. Needs and subjective well-being around the world. Journal of personality and social psychology 101, 2 (2011).
Tubadji et al. (2020) A. Tubadji, F. Boy, and D. J. Webber. 2020. Narrative economics, public policy and mental health. Covid Economics 20 (2020).
Usher et al. (2020) K. Usher, N. Bhullar, and D. Jackson. 2020. Life in the pandemic: Social isolation and mental health. Journal of Clinical Nursing (2020).
van Lenthe et al. (2015) F. J. van Lenthe, T. Jansen, and C.B.M. Kamphuis. 2015. Understanding socio-economic inequalities in food choice behaviour: can Maslow’s pyramid help? British Journal of Nutrition 113, 7 (2015).
Wahba and Bridwell (1976) M. A. Wahba and L. G. Bridwell. 1976. Maslow reconsidered: A review of research on the need hierarchy theory. Organ. Behav. Hum. Decis. Process 15, 2 (1976).
Weber et al. (2012) I. Weber, V. R. K. Garimella, and E. Borra. 2012. Mining web query logs to analyze political issues. In WebSci.
Wellenius et al. (2020) G. A. Wellenius, S. Vispute, V. Espinosa, A. Fabrikant, T. C. Tsai, J. Hennessy, B. Williams, K. Gadepalli, A. Boulange, A. Pearce, et al. 2020. Impacts of state-level policies on social distancing in the united states using aggregated mobility data during the covid-19 pandemic. arXiv:2004.10172
West et al. (2013a) R. West, R. W. White, and E. Horvitz. 2013a. From cookies to cooks: Insights on dietary patterns via analysis of web usage logs. In WWW.
West et al. (2013b) R. West, R. W.. White, and E. Horvitz. 2013b. Here and there: Goals, activities, and predictions about location from geotagged queries. In ACM SIGIR.
Yang and Li (2013) H. Yang and Y. Li. 2013. Identifying user needs from social media. IBM Research Division, San Jose (2013).
Zhang and Zhou (2013) M-L Zhang and Z-H Zhou. 2013. A review on multi-label learning algorithms. IEEE Trans Knowl Data Eng 26, 8 (2013).

Appendix A Online Appendix

A.1. Analyzing National Representation

To understand how much of the US population is represented by the collected data, we obtained demographics data from the Census Reporter API (Bureau, 2018) for ZIP codes represented in our dataset. Census Reporter API provides demographics data for 32,989 US ZIP codes, and not all of the ZIP codes in our dataset have available demographics information through this service. Between our dataset and the demographics data, we have 96.4% overlap of ZIP codes, representing 97.5% of total queries in our dataset. Table A1 summarizes the median of the select 11 demographic variables for ZIP codes in our dataset in comparison to all available ZIP codes in Census Reporter. Given 96.4% overlap, we find that the ZIP codes in our dataset closely mirrors the US population. Although query volumes are not uniformly distributed across these ZIP codes, the vast majority of US ZIP codes is included in our dataset.

	Our Dataset	Census
Population	3,072	2,775
Median Income	$54,231	$54,048
Median Age	41.8	41.9
% Race White	0.87	0.88
% Male	0.5	0.5
% HS Grad or Higher	0.9	0.9
Gini Index	0.42	0.42
% Below Poverty Lvl.	0.12	0.12
% Housing Owned	0.76	0.76
% Has Internet	0.77	0.76

Table A1. Distribution of demographic variables (median) for US ZIP codes in our dataset compared to all of the available US ZIP codes in the census data.

A.2. Analyzing Selection Bias

We sought to understand potential biases in socioeconomic circumstances that would influence the usage of Bing search engine. We leveraged deidentified client id as a proxy for a unique user to estimate the ‘client rate’, or how much of the population in a ZIP code is using the Bing search engine. We examined the correlation between the client rate and various demographic factors at the ZIP code level (e.g., income, race, age, gender, education, housing, internet access). Although unsurprisingly the factors describe some of the variance, none were correlated more strongly than $r$ =-0.058 (% Housing Owned). Table A2 summarizes Pearson correlation statistics. These results suggest that our dataset is not strongly biased towards any single demographic.

Demographic Variable	Corr Coeff
% Housing Owned	-0.058***
% Female	-0.029***
% Race White	-0.024***
Median Income	0.021***
% Has Internet	0.019***
% HS Grad or Higher	0.018***
Median Age	-0.010
% Below Poverty Level	-0.007
* $p<0.05$ , $p<0.01$ , * $p<0.001$

Table A2. Pearson correlation between Bing client rate and demographic variables for each ZIP code.

A.3. Reliability of Search Interaction Trends

Although our analysis relies solely on Bing search data, many Americans use other search engines such as Google. Therefore, we compared search trends for Bing with data available via the Google Trends API⁶⁶6https://trends.google.com/ for the same time period and a subset of keywords. We chose to use select keywords from each need category, rather than applying our needs aggregation pipeline (described in Sec. 4.2), because the Google Trends public API does not support regular expressions or access to the click interactions. We chose 9 keywords that were representative of their respective need subcategory (i.e., ‘online learning’ captures online education queries), had a significantly larger query volume compared to other keywords in the need categories (i.e., ‘hand sanitizer’ had more query volume than ‘mask’), or may have a seasonal effect (i.e., query volume for ‘tax’ could depend on the tax season). We conducted a correlation analysis on a moving average of a full week to account for timezone differences between the two data sources. Visual inspection of both Google and Bing trends confirm that search patterns across these two search engines are remarkably similar (Fig. A1), and Pearson correlation coefficients are very high with a median of 0.96 (min=0.45, max=0.98, all $p$ ¡0.001). Table A3 summarizes Pearson correlation outputs. These results imply that our findings are not simply an artifact of using one search engine over another.

Need	Keyword	Corr Coeff
SA	recipe	0.960***
SA	netflix	0.935***
SA	“online games with friends”	0.896***
Cog	“online learning”	0.973***
L&B	“online dating”	0.448***
Safe	unemployment	0.977***
Safe	“hand sanitizer”	0.966***
Safe	tax	0.913***
Safe	gun	0.764***
Phys	“grocery delivery”	0.980***
Phys	coronavirus	0.964***
Phys	“food stamp”	0.963***
Phys	health	0.888***
* $p<0.05$ , $p<0.01$ , * $p<0.001$

Table A3. Pearson correlation between Bing search trend and Google search trend for each keyword. Quotation marks indicate that the entire string is matched.

A.4. Human Needs Categories and Detection

We draw inspiration from Maslow’s hierarchy of fundamental human needs (Maslow, 1943, 1970) to tag each search interaction. Of the eight top-level human needs from Maslow’s expanded hierarchy of needs, we omit Esteem, Aesthetics, and Transcendence because they are difficult to operationalize from observational data alone. We categorize search behaviors into one or more of five broad categories of needs. Safety and Physiological are considered as ‘basic’ needs. Love and Belonging, Cognitive, and Self Actualization are often considered to be ‘psychological’ needs, and Cognitive and Self Actualization are considered as ‘growth’ needs.

We obtain further granularity of the 5 human need categories by decomposing each need category into many sub-need categories. This extra level of granularity has advantages in teasing apart the specific aspects of the need (e.g., shelter vs. finances in Safety needs) and in debugging and refining the detection logic. Table LABEL:tab:suppsubneeds enumerates 79 sub-needs and their examples.

For each need subcategory, we had multiple regular expressions applied to either the query string, the clicked URL, or both. This is noted in Table LABEL:tab:suppsubneeds under the ‘Logic’ column. ‘Keyword and domain (KD)’ logic indicates that there is one regular expression for matching query strings and another regular expression for matching clicked URLs. Both must be matched for the search interaction to be categorized as that need. ‘Queries (Q)’ logic indicates that there is one regular expression for query strings alone regardless of the clicked URLs. ‘Domains (D)’ logic indicates that there is one regular expression for clicked URLs regardless of the query strings. A full table of regular expressions used in the data collection is provided in a separate supplementary file and can be cross-referenced through the ‘Need Id’ column⁷⁷7https://github.com/jinasuh/pandemic_needs.

There were several steps in arriving at these subcategories and regular expressions. We first identified a set of e-commerce websites (1) from market research blog posts and reports that were publicly available online and (2) from enumerating top 100 URL hosts from a sample of search interactions that were tagged as purchase or commerce related using built-in classifiers in Bing. We combined these two data sources to manually curate a list of 61 e-commerce domains.

Next, we collected another sample of query strings that led to subsequent clicks to these e-commerce domains as well as page snippets that were displayed on the search result page. Using the Latent Dirichlet Allocation module in gensim, a popular topic modeling tool in python, we trained an unsupervised topic model on the query strings and the associated page snippets for 100, 500, and 1000 topics on 100,000 unigrams. We manually inspected the keywords in the topics to curate 22 topics of over 800 keywords. We also extracted frequent bigrams from the same dataset to expand our keywords if existing unigrams were too ambiguous. These topics represented purchase categories because they were collected from search interactions with subsequent clicks to e-commerce websites.

Our next step was to categorize these topics into five main need categories. We used a hybrid card sorting method (Albert and Tullis, 2013) with 5 participants who merged, split, edited, or created these topics into need subcategories. Then these need subcategories were then placed into the 5 main need categories.

Using the outputs from the card sorting activity as a basis, we further brainstormed subcategories beyond purchase categories. Several researchers independently brainstormed and categorized subcategories which were combined and resolved collaboratively through consensus meetings. We referred to the definitions presented in the theories of human needs to resolve any remaining disagreements (Maslow, 1943).

Once the need categories and subcategories were identified, we leveraged keywords and domains from manually curated topics and card sort outputs. We also obtained a list of topical domains from publicly available blog posts and articles summarizing recommended websites (e.g., search for best parenting website) or manually curated a list of domains (e.g., search for unemployment benefit page for each state). Again, we independently brainstormed and curated lists of keywords, keyword patterns, and domains for each subcategory and collaboratively iterated on the lists through consensus meetings. These keywords, keyword patterns, and domain URLs were combined into many regular expressions that we used to tag each search interaction.

We validated our detection logic by collecting human labels from Amazon’s Mechanical Turk, and our evaluation set achieved a precision of 97.2%, as described in detail in Sec. 4.3. A screenshot of an example task seen by a crowd worker is shown in Fig. A2, and the full set of detailed task instructions can be found at the end of this Appendix.

Table A4. Human need categories and subcategories with examples. Under the Logic column, ‘KD’ refers to matching both query string and clicked URL, ‘Q’ refers to matching query string only, and ‘D’ refers to matching clicked URL only. Need Id is provided to cross-reference with a full regular expression table provided in a separate file.

Need Category	Need Id	Subcategory	Logic	Example
SelfActualization	SA1	Audio/video related purchase	KD	home theater + e-commerce url click
	SA2	Books and reading related purchase	KD	comics + e-commerce url click
	SA3	Charity related queries	Q	charity; donations
	SA4	Cooking related purchase	KD	cookbooks + e-commerce url click
	SA5	Cooking related queries	Q	how to saute onions
	SA6	Cooking site visits	D	click on foodnetwork.com
	SA7	Crafts instruction queries	Q	how to embroider
	SA8	Crafts related purchase	KD	scrapbooking + e-commerce url click
	SA9	DIY site visits	D	click on instructables.com
	SA10	Home improvement purchase	KD	circular saw + e-commerce url click
	SA11	Home improvement related queries	Q	remodeling; wood working
	SA12	Home improvement site visits	D	click on homedepot.com
	SA13	Life goal related queries	Q	goals for living; personal goals
	SA14	Media related purchase	KD	dvd player + e-commerce url click
	SA15	Music instruction queries	Q	learn to play guitar; how to play a drum
	SA16	Musical instrument related purchase	KD	piano + e-commerce url click
	SA17	Outdoor related purchase	KD	hiking shoes + e-commerce url click
	SA18	Outdoor related queries	Q	best national parks
	SA19	Outdoor site visits	D	click on rei.com
	SA20	Parenting site visits	D	click on parents.com
	SA21	Pet related purchase	KD	ferret + e-commerce url click
	SA22	Pet related queries	Q	care for frogs
	SA23	Pet related site visits	D	click on petco.com
	SA24	Photography related purchase	KD	photography + e-commerce url click
	SA25	Photography related queries	Q	how to take pictures of a bird
	SA26	Sports related purchase	KD	elliptical trainer + e-commerce url click
	SA27	Streaming media site visits	D	click on netflix.com
	SA28	Technology related purchase	KD	ethernet + e-commerce url click
	SA29	Toys and gaming purchase	KD	xbox + e-commerce url click
	SA30	Toys and gaming site visits	D	click on twitch.tv
	SA31	Wedding related purchase	KD	wedding decorations + e-commerce url click
	SA32	Wedding site visits	D	click on theknot.com
Cognitive	C1	Cognition related queries	Q	improve memory; cant pay attention
	C2	Educational degree related queries	Q	degree program; online diploma
	C3	Educational site visits	D	click on coursera.org
	C4	Online education queries	Q	learn remotely; lesson plans
LoveBelonging	LB1	Dating related queries	Q	online dating; relationship advice
	LB2	Dating site visits	D	click on tinder.com
	LB3	Divorce related queries	Q	divorce lawyer; file for divorce
	LB4	Mental health experiential queries	Q	im alone; i feel depressed
	LB5	Mental health resource site visits	D	click on talkspace.com
	LB6	Online social activities queries	Q	online activities with family
	LB7	Social network site visits	D	click on facebook.com
	LB8	Social technology uses	KD	families + click on whatsapp.com
Safety	S1	Bank related queries	Q	banks; banking
	S2	Bankruptcy related queries	Q	foreclosure;bankruptcy
	S3	COVID-19 protection purchase	KD	disposable masks; sanitizers
	S4	Domestic violence queries	Q	domestic assult; abusive relationship
	S5	Eviction related queries	Q	evicted; rent moratorium
	S6	Financial loan related queries	Q	borrower; mortgage
	S7	Firearm purchase	KD	glock holster + e-commerce url click
	S8	Housing related queries	Q	best neighborhoods; housing safety
	S9	Job search related queries	Q	job application; resume
	S10	Job search site visits	D	click on indeed.com
	S11	Rental related queries	Q	rental apartments; houses for rent
	S12	State unemployment site visits	D	click on www.michigan.gov/uia
	S13	Stimulus related queries	Q	relief fund; loan forgiveness
	S14	Tax related queries	Q	irs; tax
	S15	Unemployment related queries	Q	im unemployed; jobless benefits
Physiological	P1	Apparel purchase	KD	athletic wear + e-commerce url click
	P2	Automobile related purchase	KD	motorcycle + e-commerce url click
	P3	Beverage purchase	KD	coffee + e-commerce url click
	P4	Cookware purchase	KD	cookie sheet + e-commerce url click
	P5	Food and beverage related queries	KD	applesauce; noodle soup
	P6	Food assistance related queries	Q	food stamps; snap program
	P7	Food delivery related queries	Q	grocery delivery; deliver meal
	P8	Food delivery site visits	D	click on instacart.com
	P9	Grocery related queries	Q	grocery; groceries
	P10	Grocery site visits	D	click on albertsons.com
	P11	Health condition related queries	Q	arthritis; diabetes
	P12	Health first aid purchase	KD	bandaid; wound closure
	P13	Health measurement equipment purchase	KD	oximeter + e-commerce site visits
	P14	Health symptom related queries	Q	i lost hearing; blurred vision
	P15	Household good purchase	KD	cleaning supplies + e-commerce url click
	P16	Insomnia related queries	Q	I cant sleep; help falling asleep
	P17	Prescription related queries	Q	medication interactions; pharmacies
	P18	Prescription site visits	D	click on rxlist.com
	P19	Sleep aid purchase	KD	sleep supplement; melatonin
	P20	Toilet paper purchase	KD	toilet paper; cottenelle

A.5. Significant Changes in Human Needs

See pages - of figures/Supp-Mturk-Instructions-Doc.pdf

Population-Scale Study of Human Needs During the COVID-19 Pandemic: Analysis and Implications