This is a collection of selected projects I've worked on during my professional career so far, as well as a few Python data analysis projects. Just before Elliott Davis, I had no prior experience in Power BI or any business intelligence development, and it has been wonderful to aid clients in discovering more about their data while also building aesthetically pleasing frontend dashboards to present insights. All PDFs are exports of .pbix files I've worked on, and the data is scrubbed/modified to protect client identities. See below for brief project summaries and check out each folder's README for more details. Thank you!
- Analytics Professional Work Portfolio
- Links
- Table of Contents
- Projects Overview
- Airlines Delays Analysis
- Airlines Delays Power BI ML Project
- Automated Medical Financial Reporting
- Bing News Analysis Project
- Client Proposal Dashboard
- Current Client Analytics Demo
- Cybersecurity Analytics Report
- Determinants of Earnings Analysis
- External Webinar Dashboards
- Fatal Force Analysis
- Group Case Study Demo
- Healthcare Ongoing Dashboard
- Healthcare Private Practice
- Healthcare Proposal Dashboard
- Internal Use Work Tracker
- Joshua Project Analytics
- LEGO Rebrickable Analytics
- Manufacturing Client Demo
- Manufacturing QuickBooks Analysis
- Manufacturing Supervisor Dashboards
- Medical Billing Service Provider Dashboards
- ML Project Tutorial
- Penguins Analysis
- Power BI + Python + Machine Learning Forecasting
- Private Equity Financial Reporting Automation
- QBO Payroll Import Automation
- Real Estate Client Dashboards
- Recreation.gov Analytics
- Space Missions Analysis
- Other Related Work
- Author
- Acknowledgments
Python data science project I came up with as a capstone to finish off Jose Portilla's course Python for Data Science and Machine Learning Bootcamp on Udemy. This is a portfolio project to showcase what I learned from the course. The five part analysis and series of machine learning models digs into the Airlines Delays dataset from Kaggle. Loads data in, cleans as necessary for NaN or null values, and has a focused scope for each notebook.
Beyond client projects and engagements I wanted to spend some time learning on my own other features Power BI has and what else it's capable of, and I stumbled across a couple of awesome tutorials from Microsoft and TheOyinbooke on YouTube, Github link here. This is a mini project that combines aspects of the ML tutorial project and my Python airlines delays analysis.
This was a great challenge with my manager to develop a solution for a client who had a manual and very time-consuming reporting process and needed something more efficient - not just for automating, but also making Power BI look less like a traditional dashboard and more like a printed report. We were able to help this client cut down their time from 1.5 weeks to about 1 day of making this report. This project was also featured in External Webinar Dashboards and Case Study and Blog Posts.
As part of preparing and learning for Microsoft's DP-600 certification, Fabric Analytics Engineer Associate, I knew I wanted to get practical project experience. I am really happy to have found this complete end-to-end project course on Udemy. Mr. K walks you through a complete data and analytics engineering project from start to finish, constructing a sustainable pipeline from data source to report, and even adds in alerts at the end with overall testing procedures. I wanted to highlight this starter project and other personal Fabric work to showcase what I'm learning and exciting capabilities of analytics engineering.
As my director was planning for potential engagements, he was looking for a template demo dashboard we could have on hand which we could easily customize for proposals. One of my colleagues started this one and I helped fine-tune the data model and take it further visually to produce the result seen here. The key idea is not only to display how we could personalize Power BI reports towards a potential client, but also showcase the basic yet powerful capabilities in these reports - from filtering, to visual interactivity, and all the way to maps and time trends.
After working through Healthcare Practice Demos, I was able to borrow quite of a few of the new functionalities I developed to make a potential addendum dashboard series for an existing Elliott Davis client. The main interest was analyzing how loans were paid off over time, and being able to slice and dice categories by different measures. It was satisfying being able to test different visuals and how they would work best given the provided data and what needed to be seen.
Another division of the Digital Consulting Practice Group, Cyber & Penetration Testing, had an interest in finding out how analytics could increase their value in service offerings. I was brought on to develop a potential analytics report that the Cyber team could give to their clients before they begin a security engagement as a way to set the stage for their understanding how secure they are, as well as existing vulnerabilities.
Python data science project from Anglea Yu's course 100 Days of Code: Python on Udemy. This is a professional portfolio project to showcase what I learned from the 100 day challenge. The analysis digs into the Determinants of Earnings Datasets from Kaggle. Loads data in, cleans as necessary for NaN or null values, and explores the data via a series of questions.
On April 27, 2023 I was a panelist for an external webinar Elliott Davis hosted about data analytics and how businesses could leverage data in their workplaces. The dashboards in this set are the ones I screenshared during the event, while three of my colleagues voiced-over and talked about the business cases, what we did to help solve, as well as deliver a quick demo of how Power BI works. It was a huge success and one of the firm's most popular webinars to date, and helped garner much more interest in analytics and discovering the power behind their data.
Python data science project from Anglea Yu's course 100 Days of Code: Python on Udemy. This is a professional portfolio project to showcase what I learned from the 100 day challenge.
This analysis digs into the following datasets from Kaggle:
- Fatal Police Shootings in the US
- Percentage of People Below Poverty Level
- Police Killings in the US
- Sharing Race by City
Loads data in, cleans as necessary for NaN or null values, and explores the data via a series of questions.
Earlier on in my time at Elliott Davis my manager was looking for a dashboard pack we could showcase in a variety of settings to get the word out on the Data & Analytics service line at our firm, as well as provide a starter point of conversation for clients interested in adding on analytics for their engagements. A couple of colleagues and I first worked on our own files and individual dummy data models across different industries - and then from which I combined everything into one file we could post online to our Power BI Service Workspace. As with other demos/proposals I worked on this one was a great showcase of what my team and I could do as well as the significant functionality within Power BI.
I was super encouraged when one of the first projects I worked on became a monthly addition to an existing client. My manager and I worked on developing this one together for a couple of months, fine-tuning it with the current engagement team, and this client loved it enough to make an addendum to subscribe for monthly updates on their data. As of now I am maintaining the production version of this file and providing analytics updates for them each month.
These are a set of Power BI demos intended to go to market for existing Elliott Davis healthcare clients. My director had experience in Tableau so some of the functionality has was used to was not natively built-in with this app, so this was a great challenge to push myself with what Power BI is capable of technically in order to provide a better solution. I also tested this project with a bit of Python scripting.
Instead of a generic demo dashboard, my manager needed a set of dashboard reports to show to a healthcare client in an effort to set up an engagement. We borrowed from existing datasets given to us from this client in order to create what is seen here. From initial conversations they were interested in seeing how their different service lines performed by year across different categories, as well as find out if their current Excel reporting package could be automated in Power BI or be made more efficiently in general. I focused on these ideas when working on this one.
This was a great challenge because instead of creating dashboards for clients, now I was tasked with helping to develop a work tracker for internal use. Two of my colleagues worked on this one before I hopped on, and from their work I helped communicate with a few members of upper level leadership to bring it home. The overall idea was to gradually replace the existing solution to track work over time that was housed on one platform, and move it into Power BI - with the goal of course to create something even better. Many people from different service lines and divisions would be using this tool so the end result needed to be flexible but specific enough for their needs.
As part of continuing to hone in my data and analytics engineering skills, as well as become more familiar and skilled in Fabric, I thought of other topics I'm passionate about and came across Joshua Project's public API. They are an organization keeping track and reporting on where Christians are around the world, specifically focusing on countries that have not heard the gospel or have Bible translations. Their data source is rich with tons of valuable of information, and it was difficult narrowing in on what to show for the resulting reports! Working on a plausible analytics engineering solution for Joshua Project was a lot of fun and super valuable to continue learning and growing, as well as share a topic I'm passionate about.
I was really excited to dig into this project and fine-tune my analytics engineering skills because Rebrickable has public LEGO data available, and LEGO has always been one of my favorite hobbies ever. They inspire creativity and imagination, remain high quality with the system they developed throughout the years, and I always loved the different themes and storyline to get lost in as a child. It's encouraging for me that many more people are discovering LEGO for themselves and building models from sets or on their own with the parts they have. I've used Rebrickable quite a bit to discover other creators' custom MOCs (my own creations) and love that they have these publicly available datasets across many different entities, such as colors, themes, sets, and categories. I wanted to have fun with this one and practice my analytics skills with a scalable end-to-end solution for LEGO Rebrickable data.
It's always interesting to discover what new challenge you might come across when it comes to a Power BI project - no two are the same. The main goal for this engagement was to redesign and build new reports from the client's existing Power BI environment, so there was less data modeling and engineering involved as much as a greater focus on what reports they currently had available and what other views/functionalities they would like to see. I'm thankful for this one to continue practicing making an effective and beautiful user experience as well as effectively communicating feedback with clients for delivering a better product and service.
One of our current clients liked what we built for them so much that they wanted us to focus on their financials to produce a new set of reports. This proved to be a great challenge in multiple ways - from connecting to more than one instance of QuickBooks, discovering the best way to bring the data they need in, and deepening our relationship and trust between us and the client. I also enjoyed figuring out how to cleverly manipulate DAX in different ways to achieve certain view they wished to see, including switching between MTD and QTD on the overview page, as well as the summary view table at the end. These new reports gave our client greater insight and transparency into their data to make stronger financial decisions.
After working on several healthcare and private equity financial reports, it felt great delving into the operational side of a business, specifically in the manufacturing industry. For this one I was able to help our client explore and understand more from their existing data, drawing connections and insights they haven't seen before to better understand how they can train and equip their staff. This dashboard series explored questions like: "Are my employees improving over time?" "How are our different markets stack up against each other?" "Who among a supervisor's employees might need further training?"
This was a fun one for many reasons: this client approached us in need of automating a handful of reports they already used, as well as developing a set of brand new ones that would help them clean up their business processes and discover new insights. It was also fantastic experience working on another healthcare industry-related project and finding out more about their business in each week we collaborated. It was very much a partnership between us from the beginning - making sure we understood their math & logic to work through the DAX challenges we faced, as well as diving into their forecasting.
Beyond client projects and engagements I wanted to spend some time learning on my own other features Power BI has and what else it's capable of, and I stumbled across a couple of awesome tutorials from Microsoft and TheOyinbooke on YouTube, Github link here. This tutorial is a mini project that explored a few of the AI functions behind the scenes in Power Query as well as standing up a full ML model in Power BI Service's Power Query.
Python data science project I came up with as a capstone to finish off Jose Portilla's course Python for Data Science and Machine Learning Bootcamp on Udemy. This is a portfolio project to showcase what I learned from the course. The five part analysis and series of machine learning models digs into the Penguins dataset from Kaggle. Loads data in, cleans as necessary for NaN or null values, and has a focused scope for each notebook.
My director asked one of my colleagues and I to try building a machine learning model to predict revenue, using either R or Python. The key idea was to spend time investing in research for building a scalable model that could be used for potential future engagements with clients to add greater value. This was a great challenge not only putting together concepts I taught myself in an online Python machine learning course, but also think strategically about our work and what we could help customers unlock with their information for better decision making.
In a similar vein as Automated Medical Financial Reporting, this one was another good challenge in transforming a client's existing financial reporting deck into a fully automated Power BI report. I took over this project half-way in and helped finish it up week by week, ensuring that the math and design matched what they needed. There were also some tricky formatting needs they had that was a good test to push myself in what I can develop in Power BI.
There was an internal team whose work with a client had time-consuming processes to cleanse and prepare QuickBooks Online data. I was able to step in and help them save time on a few, including this preparing this payroll import, by using Power BI and Power Query.
There was an existing real estate tax client of the firm that was interested in a reporting solution for a few of their KPIs, and my team was able to step in and help. After exploring several options of what we could do, we decided on building reports with a new platform we had not used yet, Google Looker, to increase our analytics and reporting capabilities.
I wanted to get more hands-on Fabric analytics engineering project work, and thought it would be fun to create solutions for topics I'm passionate about. Nature is one of the best gifts we have, and I'm grateful for all the amazing ways we can get out there and explore, discover new trails and enjoy the world around us. Hiking mountains is one of my favorite things! In 2023 I actually used Recreation.gov to book a ticket for Multnomah Falls in Oregon, not knowing that they actually had a public API I could use with the account I made. For this project I dug into their documentation and explored their data, and built an concept analytics solution that focuses on how someone at the organization could funnel data into a stable and scalable data model, and then build reports they could share either with the public or use internally.
Python data science project from Anglea Yu's course 100 Days of Code: Python on Udemy. This is a professional portfolio project to showcase what I learned from the 100 day challenge. The analysis digs into the Space Missions Launches dataset from Kaggle. Loads data in, cleans as necessary for NaN or null values, and explores the data via a series of questions.
Also check out a case study and a series of blog posts I contributed to Elliott Davis's Data & Analytics service line
- Case Study: Power BI Automated Reporting
- Blog Series: Power BI Best Practices - Data Modeling and Sources
- Blog Series: Power BI Best Practices - Data Engineering and Data Analysis Expressions
- Blog Series: Power BI Best Practices - Reports and Deployment
- Blog Series: Power BI Best Practices - Security and Administration
- Blog Series: Power BI Best Practices - Business Guidelines for Power BI Adoption
- LinkedIn - Garrett Becker
Thank you to my colleagues Alek Bevensee, Andrew Calandra, Nancy Thao, Brad Northington, Ellis Millwood, and many others for partnering on these projects!