Perform an analysis on Kickstarter data to uncover trends that make crowdfunding campaigns successful.
Excel is a widely used data analytical tool because of its flexibility and ease of use. It is a highly versatile tool still. Benefits of excel include the ability to use advanced formulas, create interactive charts with graphs and pivot charts and organize data by using filters and conditional formatting. Formulas and filters allow us to visualize data in order to reveal trends.
The purpose of this analysis is to examine trends within the kickstarter campaign industry, specifically for theaters and plays. An up and coming playwright wants to start a crowd funding campaign to help fund her play. She's estimating a budget of over $10,000 but wants to first see if that is an accurate estimate before jumping in to her first crowdfunding campaing. We can use excel to organize, sort, and analyze data to help her get a better sense of the accuracy of her project's estimate. We will analyze to see if there are specific factors that make a campaign successful. Using excel to analyze this data will help our client gather a better understanding of campaigns from start to finish to better plan her own campaign.
With Excel we can:
- Import data into a table for analysis.
- Apply filters, conditional formatting, and formulas.
- Generate and interpret pivot tables.
- Calculate summary statistics such as measures of central tendency, standard deviation, and variance.
- Characterize data to identify outliers in datasets.
- Perform an Excel analysis with visualizations.
- Interpret common Excel visualizations
The analysis below was performed by creating a pivot table from the main Kickstarter Data Sheet. A pivot table is a helpful analysis for this data because it summarizes it in an easily viewable filtered way.
The data became filtered down to focus only on theater campaigns at different times of year. Launch dates became the row while outcomes became the column and the number of outcomes was treated as a value.
In examining Figure 1: Theater Outcome versus Launch, we can view categories of outcomes related to time of year the campaigns lauched. Looking at the chart below, it appears that most successful theater campaigns occured when they were kickstarted in Springtime, especially May.
Logistically this makes sense, as it is likely that the general population tends to be very busy in the fall months leading up to holidays and likely don't put a lot of time or concern into booking and attending theater productions. Theater productions likely see heights of activity during summer and vacation months when families have more time for freedom, the weather is nicer, and they aren't looking to save their money away for the holidays and gifts, etc.
When examining campaigns that failed, the numbers are pretty steady across the year and there doesn't seem to be a strong trend line one way or the other in favor of a certain time of year for when a campaign will decide to cancel.
In examining Figure 2: Theater Outcomes versus Goals, we examine percentage of success to failure based on various monetary goal ranges (by the thousands). In looking at the chart below, it appears that the plays that turned out to be the most successful, either had extremely feasible goals (and therefore likely appealed to the general masses who will select based on budgetary means), or they had fairly lofty goals between 35,000 to 45,000.
Most campaigns that had goals over 45,000 did not fair so well. In comparing this trendline to percentage of campaigns failed, almost the opposite can be said, meaning that as campaigns start out with more reasonable income goals, the percentage failing is rather low and steadily increases as the campaign goals increase. There are some outliers however in the ranges 35,000 to 45,000 that cannot really be analyzed with this particular data.
- Time of year matters for launching a campaign, especially in regards to theatrical plays.
- Campaigns that see the most success tend to start in early springtime, peaking in May and slowy wind down to seeing their lowest levels in wintertime. This makes sense as plays and theaters most often occur in summertime and so they will generate the majority of their interest in the months leading up to summer. Campaigns also saw higher levels of failure in the spring and summer months, which could be due to just the higher number of campaigns kickstarting then anyways. Campaigns also had higher failure rates in October and fall.
After analyzing the dataset, we can conclude that campaigns for theatrical plays will have a corresponding correlation with the level of monetary goals set. Campaigns that had more achievable goals tended to have higher success rates while the more increasingly lofty the moentary goals became, the higher the level of failures came and the lower the number of successful campaigns. Therefore we can conclude there is generally a correspending negative relationship between financial goals set and the success of the campaign.
There are actually quite a few limitations to this dataset. For example, when I saw the outliers in Figure 2 where the success went up between the finanacial ranges of 30,000 to 45,000 and then back down after, I couldn't help but think that we were missing some sort of demographic and possibly geographic data to explain this occurrance. It would be interesting to filter down even further by city, by type of plays (comedies, tragedies), time of year, and then demographic data of those purchasing tickets. Limitations of the first data set might be the price of the tickets, again the geographic location like the data, the type of play, the age of the ticket purchasers, and also the type of play. There is also a limitation of the data analysis that we are performing. For example, even though we can see general trend lines in the line charts, we can't actually yet perform a statistical analysis to determine how strong a correlation of the dataset we might have. For example, while we can see that it seems there is a negative correlation between financial goals set and success of a campaign, we don't yet have an actual statistical measure to say how truly strong that correlation is or not. Further analysis like a regression analysis could be used if we can plot the data on a regression line.