ayoubabozer / kickstarter

Kickstarter Data Exploration & Predicting Projects Funding Success

Kickstarter Data Exploration & Predicting Projects Funding Success

Kickstarter is a global crowdfunding platform, focused on creativity and merchandising. The company's stated mission is to "help bring creative projects to life".

This project is making simple exploration and a look around to the crowdfunding data in Kickstarter.

Then our goal will be to evaluate and explore multiple supervised machine learning models, to predict if a project funding will be successful or failed before it is launched.

I took the data from Kaggle website, it includes 378,661 Kickstarter projects.

each project has 15 features:

1. ID : project ID.
2. name : the name of the project.
3. main_category : the main category the project.
4. category : a subgroup of main_category.
5. currency : the currency of the project.
6. deadline : the deadline of the project.
7. goal : goal amount in project currency.
8. launched : the launch date for the project.
9. pledged : pledged amount in the project currency.
10. state : state is a categorical status of the project.
11. usd pledged : pledged amount in USD (conversion made by KS).
12. usd_pledged_real : pledged amount in USD (conversion made by fixer.io).
13. usd_goal_real : amount of USD the project asked for initially.
14. backers : the number of supporters that actually invested in the project.
15. country : country of origin of the project.

in this project, we can find out :

• what is the most common category of kickstarter projects.
• what is the average amount of pledged(USD), goal and backers.
• what is the accuracy score of the prediction model.
• what is the most important factor in having successful project.

i will use the must-have python packages for Data Science and Finance:

• NumPy : Allowing us to work with multidimensional arrays, and a fast numeric array computations
• Pandas : Allowing us to organize data in a tabular form, and quickly loading remote data or a .csv file.
• Sklearn : Machine Learning in Python, Simple and efficient tools for data mining and data analysis.
• Matplotlib : is a Python 2D plotting library.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

• Docker

Installing

1. Copy the project to your machine

   git clone https://github.com/ayoubabozer/kickstarter.git
   

2. Get into the Dir

   cd kickstarter
   

3. Pull & Run Docker Image

   docker run -d --rm --name jupyter -p 8888:8888 -v $PWD:/opt playniuniu/jupyter-pandas
   

4. Open the app in : http://localhost:8888

ENJOY!.