Natural Hazards

, Volume 92, Issue 2, pp 907–925 | Cite as

Coupling sentiment and human mobility in natural disasters: a Twitter-based study of the 2014 South Napa Earthquake

  • Yan Wang
  • John E. Taylor
Original Paper


Understanding population dynamics during natural disasters is important to build urban resilience in preparation for extreme events. Social media has emerged as an important source for disaster managers to identify dynamic polarity of sentiments over the course of disasters, to understand human mobility patterns, and to enhance decision making and disaster recovery efforts. Although there is a growing body of literature on sentiment and human mobility in disaster contexts, the spatiotemporal characteristics of sentiment and the relationship between sentiment and mobility over time have not been investigated in detail. This study therefore addresses this research gap and proposes a new lens to evaluate population dynamics during disasters by coupling sentiment and mobility. We collected 3.74 million geotagged tweets over 8 weeks to examine individuals’ sentiment and mobility before, during and after the M6.0 South Napa, California Earthquake in 2014. Our research results reveal that the average sentiment level decreases with the increasing intensity of the earthquake. We found that similar levels of sentiment tended to cluster in geographical space, and this spatial autocorrelation was significant over areas of different earthquake intensities. Moreover, we investigated the relationship between temporal dynamics of sentiment and mobility. We examined the trend and seasonality of the time series and found cointegration between the series. We included effects of the earthquake and built a segmented regression model to describe the time series finding that day-to-day changes in sentiment can either lead or lag daily changed mobility patterns. This study contributes a new lens to assess the dynamic process of disaster resilience unfolding over large spatial scales.


Disaster informatics Earthquakes Human mobility Sentiment Social media Twitter 



This study is supported by the National Science Foundation under Grant No. 1760645 and a Virginia Tech BioBuild IGEP grant. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation or Virginia Tech.


  1. Bai H, Yu G (2016) A Weibo-based approach to disaster informatics: incidents monitor in post-disaster situation via Weibo text negative sentiment analysis. Nat Hazards 83:1177–1196CrossRefGoogle Scholar
  2. Beigi G, Hu X, Maciejewski R, Liu H (2016) An overview of sentiment analysis in social media and its applications in disaster relief. In: Sentiment analysis and ontology engineering. Springer, pp 313–340Google Scholar
  3. Bertrand KZ, Bialik M, Virdee K, Gros A, Bar-Yam Y (2013) Sentiment in New York City: a high resolution spatial and temporal view. arXiv preprint arXiv:13085010
  4. Cody EM, Reagan AJ, Mitchell L, Dodds PS, Danforth CM (2015) Climate change sentiment on twitter: an unsolicited public opinion poll. PLoS ONE 10:e0136092CrossRefGoogle Scholar
  5. EERI (2014) EERI Special Earthquake Report: M 6.0 South Napa Earthquake of August 24, 2014. USA Earthquake ClearinghouseGoogle Scholar
  6. Ford JD et al (2016) Opinion: big data has big potential for applications to climate change adaptation. Proc Natl Acad Sci 113:10729–10732CrossRefGoogle Scholar
  7. Frank MR, Mitchell L, Dodds PS, Danforth CM (2013) Happiness and the patterns of life: a study of geolocated tweets. Sci Rep 3:2625. CrossRefGoogle Scholar
  8. Geary RC (1954) The contiguity ratio and statistical mapping. Inc Stat 5:115–146Google Scholar
  9. Godschalk DR (2003) Urban hazard mitigation: creating resilient cities. Nat Hazards Rev 4:136–143CrossRefGoogle Scholar
  10. Gonzalez MC, Hidalgo CA, Barabasi A-L (2008) Understanding individual human mobility patterns. Nature 453:779–782CrossRefGoogle Scholar
  11. Granger CW (1981) Some properties of time series data and their use in econometric model specification. J Econom 16:121–130CrossRefGoogle Scholar
  12. Guan X, Chen C (2014) Using social media data to understand and assess disasters. Nat Hazards 74:837–850CrossRefGoogle Scholar
  13. Kryvasheyeu Y, Chen H, Obradovich N, Moro E, Van Hentenryck P, Fowler J, Cebrian M (2016) Rapid assessment of disaster damage using social media activity. Sci Adv 2:e1500779CrossRefGoogle Scholar
  14. Lin Y-R (2014) Assessing sentiment segregation in urban communities. In: Proceedings of the 2014 international conference on social computing. ACM, p 9Google Scholar
  15. Lu X, Bengtsson L, Holme P (2012) Predictability of population displacement after the 2010 Haiti earthquake. Proc Natl Acad Sci 109:11576–11581CrossRefGoogle Scholar
  16. Metcalfe AV, Cowpertwait PSP (2009) Introductory time series with R. Springer, New York.
  17. Mitchell L, Frank MR, Harris KD, Dodds PS, Danforth CM (2013) The geography of happiness: connecting twitter sentiment and expression, demographics, and objective characteristics of place. PLoS ONE 8:e64417CrossRefGoogle Scholar
  18. Moran PA (1950) Notes on continuous stochastic phenomena. Biometrika 37:17–23CrossRefGoogle Scholar
  19. Munich RE (2017) NatCatSERVICE: number of catastrophic natural loss events worldwide 2010–2016. Accessed 17 Aug 2017
  20. Nagy A, Stamberger J (2012) Crowd sentiment detection during disasters and crises. In: Proceedings of the 9th international ISCRAM Conference, pp 1–9Google Scholar
  21. Neppalli VK, Caragea C, Squicciarini A, Tapia A, Stehle S (2017) Sentiment analysis during Hurricane Sandy in emergency response. Int J Disaster Risk Reduct 21:213–222CrossRefGoogle Scholar
  22. Nielsen FÅ (2011) A new ANEW: evaluation of a word list for sentiment analysis in microblogs. arXiv preprint arXiv:11032903
  23. Pavalanathan U, Eisenstein J (2015) Confounds and consequences in geotagged Twitter data. arXiv preprint arXiv:150602275
  24. Pew Research Center (2017) Social media fact sheet. Accessed 11 Apr 2017
  25. Phillips PC, Ouliaris S (1990) Asymptotic properties of residual based tests for cointegration. Econom J Econom Soc 58:165–193Google Scholar
  26. Ribeiro FN, Araújo M, Gonçalves P, Benevenuto F, Gonçalves MA (2015) SentiBench-a benchmark comparison of state-of-the-practice sentiment analysis methods. arXiv preprint arXiv:151201818
  27. Said SE, Dickey DA (1984) Testing for unit roots in autoregressive-moving average models of unknown order. Biometrika 71:599–607CrossRefGoogle Scholar
  28. Swier N, Komarniczky B, Clapperton B. (2015) Using geolocated Twitter traces to infer residence and mobility. Office for National Statistics GSS Methodology Series, 41Google Scholar
  29. Tang Z, Zhang L, Xu F, Vo H (2015) Examining the role of social media in California’s drought risk management in 2014. Nat Hazards 79:171–193CrossRefGoogle Scholar
  30. UNISDR (2017) Disaster statistics. Accessed 2 May 2017
  31. USGS (2014) M 6.0—6 km NW of American Canyon, California. Accessed 11 Apr 2017
  32. USGS (2017) M 6.0—6 km NW of American Canyon, California. Accessed 9 Apr 2017
  33. Vo B-KH, Collier N (2013) Twitter emotion analysis in earthquake situations. Int J Comput Linguist Appl 4:159–173Google Scholar
  34. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D (2002) Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther 27:299–309CrossRefGoogle Scholar
  35. Walther M, Kaisser M (2013) Geo-spatial event detection in the twitter stream. In: European Conference on Information Retrieval. Springer, pp 356–367Google Scholar
  36. Wang Q, Taylor JE (2014) Quantifying human mobility perturbation and resilience in Hurricane Sandy. PLoS ONE 9:e112608CrossRefGoogle Scholar
  37. Wang Q, Taylor JE (2015) Process map for urban-human mobility and civil infrastructure data collection using geosocial networking platforms. J Comput Civ Eng 30:04015004CrossRefGoogle Scholar
  38. Wang Q, Taylor JE (2016) Patterns and limitations of urban human mobility resilience under the influence of multiple types of natural disaster. PLoS ONE 11:e0147299CrossRefGoogle Scholar
  39. Wang B, Zhuang J (2017) Crisis information distribution on Twitter: a content analysis of tweets during Hurricane Sandy. Nat Hazards 89:1–21CrossRefGoogle Scholar
  40. Wang Z, Ye X, Tsou M-H (2016) Spatial, temporal, and content analysis of Twitter for wildfire hazards. Nat Hazards 83:523–540CrossRefGoogle Scholar
  41. Wang Y, Wang Q, Taylor JE (2017) Aggregated responses of human mobility to severe winter storms: an empirical study. PLoS ONE 12(12):e0188734. CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V., part of Springer Nature 2018

Authors and Affiliations

  1. 1.Charles E. Via, Jr. Department of Civil and Environmental EngineeringVirginia TechBlacksburgUSA
  2. 2.School of Civil and Environmental EngineeringGeorgia Institute of TechnologyAtlantaUSA

Personalised recommendations