There is data available every where we look. With crashing prices on the storage media, exponential increase in computing power and the propensity of business, however small, to collect each and every bit of information has led to deluge of data.
But how useful is the data unless it is analyzed, insights are extracted and acted upon.
If the access to the data along with the expertise still lies with data-base administrators and IT professional who build models to extract insights, there would not be true data democratization.
What is Data Democratization ?
Currently, we are middle of throes of Information Age. With so much data availability every where we look, the challenge was to provide a platform/service for layperson, with no prior coding or technical background, to analyze the same and draw insights.
That’s where I find Microsoft Power Suite of Apps bridges the gap and the public preview (at the time of writing the article) of Premium Per user SKU of Power BI provides the necessary tools for non-coder to embrace data democratization. I will detail the steps wherein I collect the innocuous data to create a Machine Learning model.
- Create a MS Form to collect data – Link to the form
- Create a Flow to update the collected data from questionnaire to upload data on Onedrive
- Create a data flow in Power BI to import the data – in Premium per user workspace (represented by a diamond sign)- in Premium per user workspace
- Train the model – the method used was binary predictive model.
The results are reproduced below :
Hope I was able to explain the method to create a Machine Learning Model and explain the journey from a simple MS-Form to valuable insights.
The use-cases of ML is endless from predicting the conversion rate of advertisement campaigns, profiling of customers to understand the probability to convert to a sale or not.
These are exciting times for data democratization, waiting and eager to hear from thoughts on this.