Inspiration
https://youtu.be/p-82YeUPQh0- Ted talk that made us interested. Data is the new oxygen needed by the world; thus, when contaminated and manipulated, it can cause the same effects as Global warming interpreted as a Global WARNING. This became my inspiration as the slightest change in data can lead to the most considerable misinterpretation and misunderstanding. Data can formulate development policies or become the largest source for gigantic frauds. Let us not underestimate the power of data. It is no more numbers. It's a lethal weapon.
What it does
It is a basic AI model that compares salary based on age and gender given in the dataset. The essence of the project is to show that AI would react differently to different subsets of data. For example, in our original excel, we had 36000 cells and nine columns when we boiled it down to 3 columns, it would give us an outcome but not necessarily a product that should be used to judge a person's calibre for a job or chances of re-offence.
How we built it
We used Google collab research as our platform, pandas for dataframe, from sklearn.tree we used DecisionTreeClassifierto to create predictions, from sklearn.model_selection we used train_test_split for testing the predictions; from sklearn.metrics we used accuracy_score for finding the accuracy of predictions and used google.colab import drive to connects colab to drive. We set plot points to gender.
Challenges we ran into
We were faced with this error "ValueError: could not convert string to float" a lot. Since it was our first project using AI we didn't how our dataset was supposed to look. We had to convert the entire csv file to numbers and the mentors helped us through it.
Accomplishments that we're proud of
As our first time doing machine learning we are proud to even have the code working. It was an intense experience and no matter what happens we are proud of ourselves.
What we learned
We learned a lot about libraries in python, how difficult it is to use change data to a format that is acceptable to the program. We strongly believe however that when judging a person every aspect of data should be available to AI not just a subset as it leads to life-altering decisions.
Log in or sign up for Devpost to join the conversation.