1 of 3

Data Collection

2 of 3

Considerations for Collecting Data

  • Your data set should somehow represent all of the people that can be affected by the work you do
  • Your method of collecting data greatly impacts the conclusions you get from it (think about leading and loaded questions; going searching for a specific answer rather than a basic fact)
  • For secondary data, can you trace it back to the source? Do you trust it?
  • Think about the safe and permissible use of data; if it is personal, has it been de-identified/anonymized?

IBM estimates “Bad Data” loses 3.1 Trillion USD annually for companies in the USA alone.

Where is all the raw information going to come from? How exactly are you going to collect it (scraping, surveys, public databases)?

3 of 3

Food For Thought

  • If you’re building a survey or model based around the study of humans and human behaviour, what makes your dataset a “good” or “bad” representation of humanity?

​

  • Different people from different parts of the world and different cultures will probably feel differently about your data collection practices; how can you try and avoid misunderstandings when collecting and storing data that relates to humans?

​

  • “Data” is just one way of transferring “information” that we like because we use computers for everything. What are some examples of effective methods of transferring information in some other format?