Representing a dataset having hundreds of data points on a single report is not an easy job. If report is displayed in tabular or matrix format, generally the approach followed is pagination or filters. But consider that if neither of it is an option, when the requirement is to comparatively analyze all these data points, grid based data representation is out of question.
The first step to achieve quantitative / statistical data analysis is by using effective visualizations. Certain examples of these visualizations are Tree Maps, Lattice Charts, Correlation matrix, Table Lens etc. These visualizations might not be available out-of-box, but by embedding graphs / charts inside a matrix, such visualizations can be created very easily.
A data set would be having lots of data points in one or more fields, and those fields would be having many axis. For example, a dataset having sales of every month over the past five years. If you try to display the same on a graph, it would be almost impossible to represent 60 data points on the same graph grouped by different years, as this would look very messy. The solution in such scenarios is to create a graph having data for one quarter of a year, and then create a matrix of such graphs. So your result would have four small sized graphs for 4 quarters (i.e. 4 columns) x 5 years (i.e. 5 rows) = 20 small graphs. If you are using a bar graph as your base graph, the visualization that would get created would be called bar plot / trellis graphs. Other examples of such plots are scatter plot matrix, histogram plot matrix etc.
The first step to achieve quantitative / statistical data analysis is by using effective visualizations. Certain examples of these visualizations are Tree Maps, Lattice Charts, Correlation matrix, Table Lens etc. These visualizations might not be available out-of-box, but by embedding graphs / charts inside a matrix, such visualizations can be created very easily.
A data set would be having lots of data points in one or more fields, and those fields would be having many axis. For example, a dataset having sales of every month over the past five years. If you try to display the same on a graph, it would be almost impossible to represent 60 data points on the same graph grouped by different years, as this would look very messy. The solution in such scenarios is to create a graph having data for one quarter of a year, and then create a matrix of such graphs. So your result would have four small sized graphs for 4 quarters (i.e. 4 columns) x 5 years (i.e. 5 rows) = 20 small graphs. If you are using a bar graph as your base graph, the visualization that would get created would be called bar plot / trellis graphs. Other examples of such plots are scatter plot matrix, histogram plot matrix etc.
Below sample visualization is an example of how you can display quantitative data i.e. huge volume of data points on the same visualization and make the data comparable as well as analyzable. You might not believe but the below visualization contains thousands of data points plotted on it, and still this data is analyzable for its intended purpose.
Article 
No comments:
Post a Comment