The purpose of this post is to present a typical sales analysis that might serve as a starting point for the task of analyzing a firm’s sales data. A sample dataset from Kaggle and the latest versions of Solr and Banana1 will be used for that purpose.
As often required, the dataset needs a bit of pre-processing, such as feature transformation or column name changes, before it can be indexed.
Graph traversal features have been introduced in Solr 6 releases. These powerful features enables Solr users to run expressions that traverses graph structures in order to introduce or extract useful information. These graph traversal features are particularly useful when data is already indexed into Solr and light graph operations are required especially on top of text search. Before proceeding, a basic knowledge of Solr and graph structures is required.
Solr traversal implementation uses Breadth First Search (BFS) to perform graph traversal which is more suitable for solving search problems than its counterpart Depth First Search (DFS). It is also possible to combine graph traversal with other search or streaming operations.
In this post, we are going to explore basic graph visualization introduced in Banana v1.7. To visualize a graph in Banana, there must be at least a collection indexed into Solr with two fields: from and to that represent the adjacency matrix or, in other words, the edges of the graph. Alternatively, two collections can be used to visualize the graph: a main collection which is configured in the dashboard settings and an additional graph collection that stores the graph matrix. The main collection will be joined with the graph collection to retrieve node labels.
TL;DR if you have a raw dataset or a data indexed into Apache Solr, a meaningful analytics dashboard that gives insights and useful graphical and tabular information can be built in minutes.