Insight Miner is a Sisense tool set that leverages advanced statistical techniques to help data analysts understand important business metrics. Insight Miner automatically looks at the impact that different variables have on a significant metric of your choosing to reveal a deeper understanding of the most meaningful factors that influence it.
Connecting Insight Miner to Sisense
Download the guide below for instructions on how to install and connect Insight Miner to Sisense.
To mine your data:
- In the Sisense Web Application, select + > New Dashboard to create a new dashboard.
- Click Select Data to create a new widget and select your target variable. Target variables must be a numeric field or a binary categorical field (i.e. containing two categories like “yes” and “no”). In this example, the Average NPS Score was selected.
- Select the Insight Miner table icon.
- Click Advanced Configuration to open the Insight Miner settings.
- From the left pane, click + to add at least one explaining variable to the widget. Explaining variables are used to measure against your target variable. Explaining variables must be numeric, dates, or categorical. Note the following:
- The algorithm does not currently support measuring numeric ID’s as explaining variables.
- For categorical data, the field can contain up to 250 categories. If a field contains more than 250 categories, the field will be ignored.
- After you have added your explaining variables, click Start Mining. Optionally, you can click Apply to save the widget before mining. An pop up window is displayed that indicates the status of mining process. This process usually takes a few seconds. The more data that is mined, the longer this process will take. Insight Miner does not currently support large data sets.
- Click Insights to explore your results.
The image below is an example results mined by Insight Miner.
- Be aware of data quality and missing values. If an explaining variable has less than 10% missing values, Sisense automatically imputes values into the field using predictive models. This will allow the algorithm to crunch the data, but may slightly distort the data’s distribution from the actual data. If a field has more than 10% missing values, Sisense will ignore the field and will not impute values for target variables.
- Insight Miner does not currently support user concurrency. Only a single user should activate the mining algorithm at a time.
Hey! Was this article helpful?
Questions? Ask the community.