• Divya Malhotra

The politics of data

As political season starts ramping up, many politicians are coming out of the woodwork and it is fascinating to watch how they are using data to gain power. Since the days of the 2008 Obama campaign, political data science has ramped up. But, as much as people talk about data in politics, very few people talk about the politics of data.


What do I mean by that? Well, anyone who has created any kind of data output , whether it be a simple excel spreadsheet or an advanced recommender system, must have realized that showing data to decision makers is as much a science project as it is a project in emotional intelligence. In my career, I have had people walk out of the room when presented with data to people almost dancing on the table because the data I was showing them presented them in a good light.


But as much energy and effort is invested in how to create good visuals, data models and data prep very few articles are dedicated to navigating the politics of presenting and disseminating data. To make matters more complex, a good data scientists are typically not salespeople and are often neither trained nor inclined to dealing with the emotional leadership that is needed in disseminating data.


Here are a few things that an analyst or data scientist can do navigate the politics of data science.


  1. Set expectations. When you are developing a model that may take you in either direction, please let everyone on the team know that findings may point in any direction. You should be honest in your analysis as much as possible and not bring your personal biases in. As you dive into data, send people a "check-in" to let them know where the analysis is heading. That way everyone knows you are just the messenger doing your job.

  2. Get a champion. Have a senior executive be your champion. Make sure this is someone who is respected and will stick by you. A champion will help you navigate political landmines and will guide you to success. If you can find someone who understands analytics ...even better!

  3. Do not drop bombs in large meetings . If there is an insight that you have derived that will prove someone wrong or will ruffle feathers or will catch someone off-guard , meet with the stakeholders (especially the executives) separately and lay out the findings first. That way you develop trust with the stakeholders and enhance your reputation in the long run.

  4. Keep Notes. As insipid as this sounds, keeping good notes is keeping documentation of people's thought process. You can refer to these notes to remind what a stakeholder said two months ago in a meeting that can lead you to understand their motivations and intent. Understanding people is key to navigating politics.


I would be interested in knowing if readers have their "tips and tricks" to navigate the politics of data.