Dakota's Website

Learning web development with this as my playground.

I'm using DigitalOcean for a small hosted Ubuntu server.

Alongside Django, Nginx and Gunicorn are handling incoming requests to view this website.

I might add React to this stack?! Need to look into it more and see how I'd be leveraging that.

This was a large project involving the onboarding of this Software as a Service (SaaS) that handled purchasing, inventory, and some accounting functions.

Near the start of the business's talks with the onboarding team, I was brought in to help start data modeling what the business would require to go to and come out of Coupa. SQL Server Agent Jobs were made to automatically run stored procedures to transform our data for delivery and their data for ingestion.

Following Agile-ish practices, there were constant tweaks made as the business figured out new data they needed or alterations to what we currently had.

Coupa's Application Programming Interface (API) wasn't the most fleshed out, so data delivery and ingestion consisted of SFTP CSV file drops and GET/POST requests in Python. I handed the SFTP using Hightouch and Python and assisted in the GET/POST with Python.

The major focuses for the data was that work orders were uploaded every minute and to do complex business and accounting logic to balance the accounting sheets each month.

Dialpad is a phone SaaS. Simple business need to get data from their API and into our data warehouse for custom alerts and reporting.

I explored their API using Postman to get a sense of what would be possible or not.

Modified in-house Python scripts to collect CSV data from their API and drop it onto a network share acting as our Data Lake.

After that was landed, the data was loaded into our data warehouse where Agent Jobs could massage the data to fit business needs.

During efforts to modernize legacy code, I was able to craft a new solution for sending out alerts for data off of machine sensors.

Every minute sensor data was loaded into the data warehouse. Custom alert rules were then compared against the raw data to transform it to be alert ready.

A SharePoint Excel sheet was used to write the custom alert rules. It was grabbed by Fivetran and slapped into the data warehouse.

Using SQL Server Reporting Services (SSRS), a report ran a small SQL snippet to get any alerts that needed to be sent out.

Sales at the company could use either the Eterprise Resource Planning (ERP) software or some 3rd party portal to enter in customer orders.

With the combination of not knowing of a more robust solution and my given direction to be more hacky than spendy, I wrote a Python script to handle this.

My largest Python project to-date, it connected to the on premise Microsoft SQL Server and the cloud MySQL database to sync customer order records.

New and updated records from both tables were loaded, sorted, and then inserted or updated to the other database.