Aspiring Data Analyst | Data Scientist
ADAPT program
+
Over the last few months, I've learned a lot about the analytics software called Alteryx. I was accepted to their ADAPT Program, which is intended for those who are impacted by COVID-19, and allowed me to get a deeper understanding of how to manipulate data to derive answers. In addition to gaining access to Alteryx for an extended amount of time, I also became Core Certified in their software.
​
After completing the Adapt program, I would say my greatest achievement was the completion of the Udacity Predictive Analytics Nanodegree. This Nanodegree program really solidified my knowledge of Alteryx, and also gave me a foundation in the data visualization software called Tableau as well. The capstone project that I completed exemplifies what I learned best, and comes in three parts.
Task 1 - Clustering
A company currently has 85 grocery stores, and up until recently, all stores used the same store format for selling their products. Having all the stores use the same format is beginning to cause problems as stores are suffering from product surpluses in some product categories and shortages in others. I was asked to provide analytical support to make decisions about store formats and inventory planning. In order to do this, I clustered stores into three categories that was decided through the predictive tools that come with Alteryx.
After creating and running my workflow, I was able to get a cluster number for each store. After that, I uploaded my results to Tableau, where I created a public workbook that displays the level of sales and which cluster each store belongs to.
Task 2 - Predicting Clusters
In this task, we learn that the store chain plans on adding 10 new stores in addition to the existing stores. Now that I have my clusters, I needed to predict which clusters the new stores would belong to as well. However, since we don't have sales data for the new stores yet, we'll have to use the demographic data to determine their format. Using Alteryx, I made a workflow with three different predictive models to perform this action.
After running this workflow, I had to decide which model would be the best for predicting the format of the new stores. Using the Model Comparison Tool, I found that the Boosted Model was my best option because it had a higher F1 score than the rest of the models. Next, I used the trained Boosted Model to score the next 10 stores and find out which cluster each new store belonged to.
Task 3 - Predicting Future Sales
The last task of the capstone project required me to use the data that I had gathered to predict using Time Series Models to find out what the sales of the existing stores and the new stores would be for the next year. The final task of this project was the most difficult in my opinion because it required three Alteryx workflows and a Tableau workbook to solve the problem. The first Alteryx workflow represents my determination of whether to use an ETS (Exponential Smoothing) Model or an ARIMA (Auto Regressive Integrated Moving Average). I ended up choosing an ETS Model because it showed a more accurate prediction than the ARIMA Model.
Below, you will find the two workflows that represent using the ETS model to predict the sales of existing stores and new stores. The first workflow is for existing stores, and the second workflow is for the new stores.
As mentioned before, the ultimate goal of this project was to predict the sales of both existing and new stores, and to present the findings in an easily digestible format. Below you will find the Tableau workbook that shows this information. This workbook is fully functioning and allows you to also find the predicted individual sales data for each month by simply hovering over the graph with your mouse.
Conclusion
The entire Udacity Nanodegree took place over the course of a month, and I learned a lot about predictive analytics from it. I hope to use the knowledge that I gained from this course into other projects in the future. If you would like to access my Nanodegree Certificate, click here or click the LinkedIn icon at the top of the page. Thank you for taking the time to read about my first big project using Alteryx and Tableau. I hope this skills come in handy in the future, and help me find a job as a Data Analyst. Have a wonderful day!