Short Python Projects

---

Project 1: Predicting blight payments

Blight violations are issued by Detroit to individuals who allow their properties to remain in a deteriorated condition. Every year, the city issues millions of dollars in fines to residents and every year, many of these fines remain unpaid. Enforcing unpaid blight fines is a costly and tedious process, so the city wants to know how to increase blight ticket compliance.

The first step in answering this question is understanding when and why a resident might fail to comply with a blight ticket. This is where predictive modeling comes in. To this end, we use a machine learning model to assess the likelihood of whether a particular ticket will be paid, using information such as street name, fine amount and disposition.

• Python code

---

Project 2: Social Network Analysis

This project uses a company's email network, in which a connection exists between employees if one has sent an email to the other. A Ridge machine learning model is used to predict the salary information of employees where this information is missing, using their place in the network together with their department and degree.

Another Ridge model is the trained to assess the likelihood that a new connection will be created between two previously unconnected employees, using a variety of indices for non-connected edges (e.g. Jaccard index) to train the model.

• Python code

---

Project 3: Document Similarity

In this project, we first assess whether two sentences are paraphrases of each other, from a collection of sentences in a document. Then, a Latent Dirichlet Allocation model is built using the gensim package, from which the 10 key topics in a long document are found, together with the 10 most important words in each topic.

• Python code