2015-04-23 Open data and bias
The article 3 Cities Using Open Data in Creative Ways to Solve Problems shows three different ways to play with data and to build interesting information at a city level. Based on that, it becomes easier to improve the life of people leaving in that city. If this data is available to people taking decisions, they can take action to fix some the issues reported on the maps and they can measure the impact after it is fixed. However, if everybody knows this data, they would probably start to change their behaviour and the data will start reflecting that change. The first issue could artificially disappear without being fixed.
That's what explains the second article Randomized experimentation. By learning from the data, machine learned models end up proposing better options to people and they both forget others options are still possible.
| <-- --> |
Xavier Dupré
|
