I found this on upwork.com. I think this was someone's take-home assessment.

In the era of IoT, car prices could be distributed across many stores, and each store is considered as a centralized node in the network with its own local data. The goal of the project is to use the dataset Cars_Data_full.xls to answer the following:

Q1. Discuss a conceptual model to predict the MRSP including significant variables, correlation, and your plan for training and testing. Q2. Build both multi-linear regression model and a MLP model to predict the MSRP of a car. Q3. Validate and interpret the model using a different number of hidden neurons and significant variables. Q4. Compare results of neural networks and linear regression, and support each model by validating the results. Q5. Make a decision and offer your recommendations. Justify your decision based on your analytics results. Q6. How would your decision change if you worked in a distributed environment? What changes do you need to make to your model (regression or MLP)? Discuss your distributed modeling strategy.