Group18 Proposal

Crime and Society : The new age of offence in India

INTRODUCTION

PROPOSAL

REPORT

POSTER

APPLICATION

BACK

Background

type content here...

About the Dataset

content place..

Objectives

As an area of analysis, crowdfunding has largely featured literature that focused more on predicting the success/failure of campaigns. However, as a field of visualization, the data has relatively been left untapped; most visualizations that exist simply show the accuracy of these prediction algorithms.

<a href="/1617t3isss608g1/File%3AProjectcat.png" class="image"><img alt="Projectcat.png" src="/1617t3isss608g1/img_auth.php/thumb/9/97/Projectcat.png/250px-Projectcat.png" width="250" height="79" srcset="/1617t3isss608g1/img_auth.php/thumb/9/97/Projectcat.png/375px-Projectcat.png 1.5x, /1617t3isss608g1/img_auth.php/thumb/9/97/Projectcat.png/500px-Projectcat.png 2x" /></a>

Our objective here is to create an interactive visualization that would help us answer what segments of users would be interested in the specific project (Health/Environmental/ Technological/ Sports/Politics, etc.) that the app launches. Usage of clustering algorithms will drive the finding of patterns on segments/users as well as predictive algorithms to find the right segment of users who would fund or endorse the project that the app is trying to publicize. Both researchers of crowdfunding and group behavior as well as people interested in starting their own campaigns can benefit from such tools as they can utilize these visualizations to make better sense of the data. Because of this emerging domain, the visualizations explored would just be the beginning of what can be an ever-increasing domain of research and analysis for this growing field.

Questions to be answered from the data are:

1. Which potential users are suitable for each project category to fund the projects?
2. Where the potential donators are coming from for the proposed projects?
3. How much potential amount of money donated for new proposed projects?

Approach

• Provide the capability to convert JSON/TXT/etc. based data files to CSV/EXCEL for further analysis.
• Provide visualization catalogue for EDA and Analysis.
• Provide the capability to perform basic data exploration to understand the levels and frequencies for categorical data; min, max, median, SD, quantiles for continuous data.
• Provide the capability to remove any outliers the user deems unfit for further analysis.
• Provide the capability to apply machine learning algorithms - unsupervised (for finding clusters/patterns) as well as supervised (for predicting logistic regression or SVMs for finding the potential new funders.
• Provide the capability to visualize the algorithms.
• Communicate the interpretations/findings/conclusion of the above work.

Expected Challenges

• A lot of data manipulation and sub-setting will be needed to make different specific plots in ggplot2/plotly in R.
• Complicated codes will be involved for building layers in ggplot2/plotly and generating interactive maps visualization in R shiny.
• There would be noises in the data that need to be determined and addressed carefully to satisfy the assumptions of the algorithms.
• Visualization for algorithms would be difficult.

References

• <a rel="nofollow" class="external free" href="http://www.investopedia.com/terms/c/crowdfunding.asp">http://www.investopedia.com/terms/c/crowdfunding.asp</a>
• <a rel="nofollow" class="external free" href="https://github.com/localytics/data-viz-challenge">https://github.com/localytics/data-viz-challenge</a>
• <a rel="nofollow" class="external free" href="http://data-viz-challenge.localytics.com/">http://data-viz-challenge.localytics.com/</a>