Spiced Projects

This repo contains all of my projects that were built as part of the Data Science course at the Spiced Academy in Berlin.

Projects include:

1. An "Animated Scatterplot" detailing world fertility rates vs. life expectancy throughout the 20th Century as well as tracking population.

2. A "Titanic Predictor" model predicting one's survival chance if they were aboard the Titanic using ticket price, passenger class, sex, and age, as well as feature engineering with a column transformer pipeline to build a logistic regression model.

3. A "Bike Predictor" linear regression model which predicts how many bikes will be rented each day for a given year in Washington D.C. The model is built using XGBoost regression.

4. A "Lyric Predicotr" language prediction model which scrapes a lyric website to collect lyrics of 5 artists, collects and studies the collections, then, in a command line interface, allows a user to input a line of text and the model attempts to predict which artist wrote the lyric.

5. Week 5 utilized AWS to host and work with data in the cloud. Unfortunately, I cannot afford to pay for the EC2 machine to stay running, and therefore, this project is shut down for the moment.

6. A "Slackbot ETL" docker pipeline which, when built and run, will set up containers that scrape a reddit page, collect the data in a MongoDB, tranfer those JSON files into a PostgresDB where the text of every post is analyzed using Vader Sentiment Analysis, and then the titles of each post and the sentiment of each post is sent to a Slackbot which posts automatically when the Docker-Compose is run.

• NOTE: I understand that this process is convoluted and not efficient, the extra steps were part of the learning process with Spiced Academy.

7. A time series forecast, which attempts to predict temperature. Steps include creating trend, seasonal, remainders, lags, and finally running linear regression to develop a prediction. Included is a pip installable function which can be used to fill in missing, or incorrect data. See: https://pypi.org/project/fill-dt-data/ for more info.

8. A Markov Chain simulation project which tracks fake customer movement data inside fake grocery stores.

9. An Image Classification project which can capture images and predict what is contained within the image. (As long as that image is a water bottle.)