{"id":7822,"date":"2018-11-12T17:22:04","date_gmt":"2018-11-12T22:22:04","guid":{"rendered":"https:\/\/www.carnaghan.com\/?p=7822"},"modified":"2019-01-04T09:45:46","modified_gmt":"2019-01-04T14:45:46","slug":"rapid-reporting-and-visualization-development-with-watson-analytics","status":"publish","type":"post","link":"https:\/\/www.carnaghan.com\/rapid-reporting-and-visualization-development-with-watson-analytics\/","title":{"rendered":"Rapid Reporting and Visualization Development with Watson Analytics"},"content":{"rendered":"
Watson<\/span> Analytics<\/span> provides a powerful suite of display and rapid reporting options. This article looks at two distinct datasets from Kaggle<\/span>, Laptop Price and Mobile Price Classification. By using the display options available within Watson<\/span> Analytics<\/span>, the ability to create a compelling story possible. Both sets of analysis are focused around the target variable<\/span> of price (or price_range), and the details of these datasets have been included below for reference. Laptop Price consists of records of various laptop models, that were last updated six months ago from writing, adding additional laptop characteristics and prices. Mobile Price Classification provides data<\/span> on mobile phones including price range classifications.\u00a0 Analysis of data<\/span> quality and cleansing will not be discussed in this article. Rather the focus of building compelling visualizations and telling a story are the priority here.<\/p>\n Watson Analytics<\/span> has been built to be accessible to non-technical users. It provides a set of tools for discovering patterns within datasets and building interactive visualizations. One of the key features of the application is Natural Language<\/span> Processing. According to Liddy (2001), Natural Language<\/span> Processing is a range of computational techniques for analyzing and representing naturally occurring texts at one or more levels of linguistic analysis for the purpose of achieving human-like language processing for a range of tasks or applications. In Watson<\/span> Analytics<\/span>, generating visualizations and reports for management is done in three main stages:<\/p>\n As mentioned earlier, no further cleansing or refinement of the data<\/span> has been done for the purposes of this paper. We will be looking primarily at the second two stages, which include discovery and display generation.<\/p>\n In this discovery set, price_euros<\/span> (or price) was the target variable<\/span> and is shown in the very first visualization<\/span> below. Several other main visualizations<\/span> were created as documented below, which focus on price drivers including a decision tree:<\/p>\n In two of the discoveries, further filtering of the data<\/span> was required to provide meaningful results. This was the case with both Price by RAM<\/span> and Price by Resolution. For Price by RAM<\/span>, the number of options of RAM<\/span> was reduced to provide a greater understanding of the impact on price. This involved filtering out lower RAM<\/span> amounts. In the dataset, the resolution value was combined with other attributes including the type screen, touch screen features, etc. In order to get a more meaningful visualization<\/span>, only those values that contained the actual resolution was used in the filtered discovery.<\/p>\n In this discovery set, price_range was the target variable<\/span>, complimented by several other visualizations<\/span> including another decision tree model<\/span>. All visualizations<\/span> have been included in Appendix 5-8. The visualizations<\/span> in this discovery include:<\/p>\n Each of the visualizations<\/span> supported the story behind those factors that drove price range. The first visualization<\/span>, Price Drivers, utilized a spiral chart to illustrate the top factors impacting price. This was later converted into a decision tree in the last visualization<\/span> described above.<\/p>\n Once the main discoveries had been created, the next step was to publish these into displays that would be suitable for management or other project stakeholders. The display options in Watson Analytics<\/span> provides various options for creating:<\/p>\n Two main dashboard<\/span> displays were created, laptop price dashboard<\/span>, and mobile price range dashboard<\/span>. Both of these dashboards were assembled using the drag and drop tools provided within Watson Analytics<\/span>.<\/p>\nDiscoveries<\/h2>\n
\n
Laptop Discovery<\/h3>\n
\n
Mobile Phone Discovery<\/h3>\n
\n
Displays and the Expert Storybook<\/h2>\n
\n