|Position:||Assistant Professor, The Information School|
Adjunct Assistant Professor, Computer Science & Engineering
Director, Data Science and Analytics Lab
University of Washington
|Contact:||joshblum[@]uw[.]edu / +1 (206) 685-8746
for encrypted correspondence: my public key
|My work focuses on the economic and social impacts of new technologies, and the development of new methods for the quantitative analysis of massive, spatio-temporal network data. Most of this work is based in developing and conflict-affected countries. I have open research positions for qualified PhD students.|
News and updates
- 10/2014: Are you a talented linux sysadmin with Hadoop/Spark experience? The DataLab is hiring!
- 10/2014: Recent/Upcoming presentations: UMD iSchool (10/21), World Bank (10/22), NEUDC (11/1), ACM-DEV (12/5)
- 07/2014: Recent/Upcoming presentations: KDD workshops on Data Science for Social Good and Emergency Response
- 04/2014: Check out the awesome lineup of speakers at the new Joint Seminar in Development Economics!
- 01/2014: Recent/Upcoming presentations: PAA, UC Berkeley, ACM-DEV, ICTD.
- show more...
Probabilistic Inference of Unknown Locations: Exploiting Collective Behavior when Individual Data is Scarce - joint with Ramkumar Chokkalingam, Vijay Gaikwad, and Sashwat Kondepudi
New sources of large-scale geospatial data can inform policy decisions ranging from disease monitoring and city planning to disaster management and humanitarian relief. However, existing methods for mining these data are not well suited to most developing country contexts where technology use is less intense and the digital traces are generally quite sparse. Here, we present a method for predicting the approximate location of a mobile phone subscriber that is more appropriate to contexts where the signal generated by each individual may be intermittent, but the collective population generates a large amount of data. This method works well when, for instance, an individual is not consistently active on the network or when the phone is off. Our model uses a nonparametric approach to probabilistically interpolate locations, and has the advantage of associating a confidence with each prediction. We test this method on a large dataset of anonymized mobile phone records from Afghanistan, and find that we can correctly predict a subscriber's unknown location in 76%-95% of cases, and that on average our predicted location is off by 0.2-1.9 kilometers.
Risk Sharing and Mobile Phones: Evidence in the Aftermath of Natural Disasters - joint with Marcel Fafchamps (Oxford) and Nathan Eagle (Santa Fe Institute)
We provide empirical evidence that an early form of "mobile money" is used to share risk. Our analysis is based on the entire universe of mobile phone-based communications over a four-year period in Rwanda, including millions of interpersonal transfers sent over the mobile phone network. Exploiting the quasi-random timing and location of natural disasters, we show that people make transfers to individuals affected by economic shocks. The magnitude of these transfers is small in absolute terms, but statistically strong. Unlike other documented forms of risk sharing, the mobile-phone based transfers are sent over large geographic distances and in response to covariate shocks. Transfers are more likely to be sent to wealthy individuals, and are sent predominantly between pairs of individuals with a strong history of reciprocal exchange. [View Video]
Calling for Better Measurement: Estimating an Individual's Wealth and Well-Being from Mobile Phone Data
We provide evidence that mobile phone records can be used to predict the socioeconomic status and other welfare indicators of individual mobile phone subscribers. Combing several terabytes of anonymized transactional mobile phone records with data collected through 2,200 phone-based interviews, we test the extent to which it is possible to predict an individual's responses to survey questions based on phone records alone. We observe significant correlations between asset ownership and a rich set of measures derived from the phone data that capture phone use, social network structure, and mobility. Simple classification methods are able to predict, with varying degrees of accuracy, whether the respondent owns assets such as radios and televisions, as well as fixed household characteristics such as access to plumbing and electricity. More modest results are obtained when attempting to predict a broader set of development indicators such as an individual's response to the question, "Have you had to pay unexpected medical bills in the past 12 months?"
We describe how large sources of geotagged data generated by mobile phones can provide fine-grained insight into internal migration. We develop and formalize the concept of inferred mobility, and compute this and other metrics on a large dataset containing the phone records of 1.5 million Rwandans over four years. Our empirical results corroborate the findings of a recent government survey that notes relatively low levels of permanent migration in Rwanda. [View Video]
Divided We Call: Disparities in Access and Use of Mobile Phones in Rwanda - joint with Nathan Eagle (Santa Fe Institute)
This paper provides quantitative evidence of disparities in mobile phone access and use in Rwanda. Our analysis leverages data collected in 2,200 field interviews, which was merged with terabytes of transaction-level call histories obtained from the mobile telecommunications operator.
Violence and Financial Decisions: Experimental Evidence from Mobile Money in Afghanistan - joint with Michael Callen (UCLA) and Tarek Ghani (UC Berkeley)
Private firms in conflict-affected countries face insecurity, corruption, poor infrastructure, and weak property rights. Disbursing employee wages is a challenge as cash-based payment systems are vulnerable to indirect costs in the form of leakage and theft. We implement a randomized field experiment in Afghanistan to test the effects of a mobile phone-based salary payment system on performance outcomes in a private firm with approximately 400 employees.
Freedom to Speak: How a Free Calling Network Affects Community Health Worker Knowledge and Productivity
We study the extent to which increased peer communication can improve the effectiveness of community health workers in Tanzania. Through a large field experiment in which roughly 8,000 health workers receive staggered access to a free mobile phone network, we measure the impact of this intervention on actual patterns of interaction and on health and welfare outcomes of workers and patients.
Savings in Conflict: The Impact of a Mobile Phone-Based Defined-Contribution Plan in Afghanistan - joint with Michael Callen (UCLA) and Tarek Ghani (UC Berkeley)
We study the extent to which a mobile phone-based defined contribution savings account can improve the financial capabilities and welfare outcomes of salaried employees at a large Afghan firm. Our research design uses a randomized controlled trial to understand whether a defined-contribution savings product can create an enduring increase in savings.
A Society of Silent Separation: The Impact of Migration on Ethnic Segregation in Estonia - joint with Ott Toomet (Tartu University)
We exploit a novel source of data to model the impact of migration and urbanization on segregation in Estonia. Analyzing the complete mobile phone records of hundreds of thousands of Estonians, we find that the ethnic composition of an individual's geographic neighborhood heavily influences the structure of the individual's phone-based network. We further find that patterns of segregation are significantly different for migrants than for the at-large population: migrants are more likely to interact with coethnics than non-migrants, but are less sensitive to the ethnic composition of their immediate neighborhood than non-migrants.