Data Prep for Machine Learning and Predictive Analytics

Data Prep for Machine Learning and Predictive Analytics Alex Ziko, Data Analyst

Agenda • Intro to the Veera suite of products • Explanation of how the data prep process works • Overview key tools used in the prep process • Demonstration of data prep on ML training data • Questions

Self-serve data prep and reporting Automated predictive modeling Cloud-based sharing platform

Data Prep for Machine Learning in Veera Construct Data prep and reporting Automated predictive modeling Photo credit: https: //hackernoon. com/machine-learning-is-the-emperor-wearing-clothes-59933 d 12 a 3 cc Cloud-based sharing platform

Prep Your Training Data Integrate Data from any Source Turn into Actionable Information Data Cleanup New Variable Creation Automated Mining and Predictive Modeling Transfer to Your Current ML Tool or Share and Present Results As Is

Veera Construct Data Prep Tools Commonly used nodes for large datasets and data warehouse integration

https: //www. complex. com/pop-culture/best-movies-on-netflix/ https: //www. businesswire. com/news/home/20161214005486/en/Amazon-Prime-Video-200 -Countries-Territories-World SOFTWARE DEMONSTRATION https: //variety. com/2016/digital/news/hulu-free-streaming-end-yahoo-1201832578/ Prepping Movie Data for Predictive Variables

Filter Your Data Filter down to the exact cohort of records you need for your training data.

Identify Duplicates & Deduplicate Duplicate records can easily be identified and handled.

Cleansing (Handle Missing Values and Substring Values) The cleanse node makes string level alterations without generating a new variable column

Transform Variables Parse-out String Values Use the Multi-variable formula feature to select keywords or choose a delimination character to focus on.

Transform Variables Multi-variable Creation Datasets can be augmented using existing variables

Script Node Use legacy scripting of various languages.

Output Your Work • Databases • Flat Files • FTP Ability • Schedule & Automate

Predictive Analytics Use Veera Predict to generate your model. Or send your newly created training data to your model creation tool

Questions

Rapid. Insight. co Download a 14 -day Free Trial m Free and Unlimited Training and Support – NO TICKETING Ask Us About Our Price Points Tell Us About Your Needs Thank You