kylegross/kickstarter-analysis

Performing analyses on Kickstarter campaign data to identify trends

Kickstarter Campaign Analysis

Project Overview

An analysis of Kickstarter campaign data to help guide Louise make informed decisions and follow best practices during campaign preparations.

Analysis + Challenges

This analysis did not present any challenges. Possible challenges I did take note of in the data was not referenced in this exercise. For example, there are special characters and additional spaces in both the "name" and "blurb" columns of the original data. In addition, none of the columns are directly defined, which may lead to some confusion and misinterpretation of the dataset. Although not a dependent variable in this analysis, the "spotlight" and "staff pick" columns contain true/false (or yes/no) data that is not translated. Although not key to Louise's analysis outcomes, it is important for analysts to understand every aspect of the original data.

Results

Two conclusions that can be drawn about the theater outcomes is that over the last few years, the number of theater campaigns has decreased dramatically (Figure 1). That said, there is a noticeable spike in campaigns in the spring and summer months of the year.

Figure 1:

Figure 1 proves that theatre campaigns are the most popular campaign type in the US.

Based on goals, the analysis proved that many of the campaigns that had greater goal amounts failed. Those with smaller goals proved to be more successful campaigns (Figure 2).

Figure 2:

Figure 2 shows that, within the category of theatre campaigns, plays are the most popular type in the US.

There are a few limitations to this dataset, most of which surround the conclusion that there just isn't enough data available. Examples include lower-level location data (city, state, zip), campaign personnell data (campaign manager, contact information), and expense details. Because of these limitations on available data, our analysis is extremely high level and does not have the flexibility to propose scheduling solutions based on available resources. A few other tables and graphs can be created from this dataset. It would be interesting to examine campaign outcomes by category, and then by country, to view the most popular campaign type accross the different locations. Another graph that can be created could focus on the cancelled campaigns, wherein the user would examine which campaign types are most often cancelled, whether there is a correlation with the time of year, and the like.

Kyle Gross, July 2020