Qualities that make a great data scientist
Observe, organize, analyze and keep learning. Most important of all, look for relationships within and without data and structure to find ways to algorithmically automate and extract value. Be willing to challenge yourself and don’t over rely on models (though your objective is to build a model!)-experiment to learn and from learning.
- Go backwards and then Forwards. Start by asking what is the objective? if you don’t have one, you are looking for a problem to solve, which is ok – the world is full of problems.
- Objective. Once you have that (ex: analyze NYC cab rides to find out when congestion pricing would yield most revenue).
- Gather data. Search relevant data sources and ask yourself if the data is reliable, accurate, and up-to-date (for the time period of your objective). Read often to get different ideas and perspectives.
- Clean data. Often, the data will need to be cleaned, structured, formatted, parsed and modified to fit into a standardized way to be analyzed evenly.
- Analyze. What might be the best ways to get to your objective? if you don’t know, experiment.
- Build the model/implement the solution.
- Outcome. Is your diagnosis / prognosis directly related to the objective? might it viewed by others as biased, skewed or incomplete? if so, go back to step 2.
- Communicate. Share your work and take constructive criticism and make improvements as needed.
Last, but not least: Take a break– do something else not related to data science 🙂
By: Hari Santanam