It was a nice experience to make kids familiar with the data science. We believe that experiments are best way to make kids realize the strengths of data.
Alpha and Beta were given 48 pictures of different faces, with accompanying names, and a favored activity. They were instructed to rate each person on the scale of 1 to 5 points whether they would want to be friends with him. The data set was further divided into 36-12 train-test set.
Then data sets were exchanged among groups. We gave the data collected from Gamma to the Alpha and Beta.
They then attempted to answer the following questions by aggregating the data and making visualizations
- Is Gamma likely to make friends quickly i.e. how many of the persons represented by the cards would he like to make friends with?
- Does Gamma have clear preferences for making friends – in terms of their gender, their name and their expressions?
Afterwards, Alpha and Beta analyzed data points and saw different plots. We also saw the distribution of features values given Gamma wants to be friend with guy/girl in the photograph. We tried to predict the likelihood that Gamma would make/not make friends with someone by putting weights on the features she considers.
It was really a nice experience to guide Alpha and Beta and help them experience how data sciences works. They were inspired to record data about their activities in such a way that they can infer meaningful information and use it in their daily lives.
Kids were really excited about it and wanted to join us again if they could learn more.
Initially, we were afraid that kids may not have got the whole experiment. But, kids were really smart and trained to align with data science approaches.
- Alpha was quickly able to infer from the graph of ratings that Gamma was very friendly guy and could decide whether to make anyone friend or not.
- Beta was good with maths, she could quickly do the maths and visualize raw data in percentages.