Blumenstock describes the potential usefulness of acquiring data on large populations through digital means. By tracking data on people through things such as mobile phones, researchers would be able to analyse and use machine-learning algorithms that would be able to predict and react to situations such as natural disasters, credit scores, poverty, and more in a much more effective and efficient way. While this would seem highly beneficial, Blumenstock discusses several “pitfalls” with using such methods of data collection. The first problem is that the solutions that are executed are susceptible to unintended consequences, due to the fact that data could possibly be misrepresenting populations. The second problem Blumenstock discusses is the credibility of algorithms and digital data that has only recently become more popular as a means of collecting information. Blumenstock notes that some of his research has involved maps that have inaccurately portrayed information over time as variables and the relationships between them change. The third problem Blumenstock discusses is how tools for research are susceptible to bias, and how groups of people that are unfairly represented are often marginalized. Data collection may unintentionally exclude groups of people, which has often led to the limited consideration of the poorer and disadvantaged populations. The fourth problem that Blumenstock discusses is the lack of regulation. In more developed countries, there are stricter policies that government agencies and aid organizations have to adhere to. However, in less developed countries, these policies are less strict, if existing at all. This ultimately leads to companies attempting to maximize profits taking advantage of the people in these developing countries. After going over several issues with digital data collection, Blumenstock suggests several ways to improve the situation. The first way Blumenstock suggests is to compare new sources of data with old sources of data, as opposed to replacing. The second way is to be able to customize the algorithms being utilized so that they can be fully effective. By ensuring that it is flexible, researchers will be more adept at considering factors such as location, people, traditions, and more. The third suggestion from Blumenstock is to expand efforts to ensure that researchers are covering a wide range of expertise and understanding. Blumenstock mentions the importance of collaboration between people of both the private and public sectors, governments, and particularly organizations that are based in the specific area that the data is involved with.