Detection of Data Errors
One of the unsung, but valuable, uses for data visualization is to find outliers and potential data entry errors. Graphs make them easy to spot. The graph on the left shows several unusual data points. For example, there are a cluster of three data points on the top left of the screen that do not follow the general trend of the data and are by themselves.
If this was a real analysis, clicking on the data points would open the record behind the information so that you can see why these points are different. You may find that these data points are critical findings, such as exceptions that could cause a life threatening situation. Or, they could be the data points that unveil a new major finding that leads to success.
On the right are pie charts showing another aspect of the same data. But, looking at the legend quickly it shows that there are some errors in the data. For example, “Hospital (Inpatient)” appears to have two entries. It is likely that one of the two entries contains an extra space. This type of coding error could throw off a typical analysis. But, the problem is easy to spot using a legend in a chart.
Practical Computer Applications (PCA) can help you determine the best back end and front end tools for your data and applications. Contact us directly at 617-527-4722 or visit the contact us page or see additional data visualization demonstrations.
Most Popular Pages
Recent Posts: Practical Blog
- Power View empowered practical Smart Client applications
- Tableau vs Qlikview and Tableau vs Spotfire
- Data, Story, View: Prototype and Update
- 2012 starts with Tableau 7.0
- SQL Server 2012 is good for Data Visualization
- Spotfire 4.0 is available
- Introduction of QlikView 11
- VisiCalc jubilee and history of Excel
- Data Visualization 2011: Omniscope 2.6
- Data Visualization 2011: Spotfire 3.3, Silver 2.0
- Data Visualization 2011: Tableau 6.1
- SQL Server 2012 is released for Preview