As part of the initial 5 weeks of training at The Data School, each Friday brings a unique challenge designed to stretch our analytical thinking, technical and consulting skills, and presentation abilities. At the start of the day, we’re given a client-style brief—often simulating real-world data problems of an industry we have no knowledge or experience with —and by the afternoon, we present our findings and visualisations to the coaches who act as the client.
The feedback we receive is detailed and constructive, focusing on both what we did well and where there’s room for growth. This process allows us to sharpen our technical skills in tools like Tableau and Power BI, and also builds our ability to communicate insights clearly and confidently, all within a time constraint.
After each project, we’re encouraged to reflect on the experience by writing a project log. This includes an overview of the task, a description of our solution, the feedback we received, and our own thoughts on what went well and what we can improve on next time.
In this blog post, I’ll walk through one of these projects—sharing my approach, the outcome, the client feedback, and the key lessons I took away.
OVERVIEW
My client was a Fire Marshall for the state of Georgia, covering Atlanta and Hapeville.
STAKEHOLDER(S)
Robbin Vernooij – Fire Marshall
BACKGROUND The Fire Marshall wanted to gain some insights to better understand the callouts that have occurred in the state of Georgia and what type of incidents require fire service resources.
OBJECTIVE
The Fire Marshall had four main decisions and insights that they needed to make:
- They needed a regional breakdown of incidents in Georgia to understand if there are any areas that need additional support
- They wanted to understand what the regional response times are in order to decide which areas need improvements to respond quickly
- They wanted to identify which incident types are more recent and which locations these are occurring in to assess how effective staff in the various fire stations respond and to identify whether improvements need to be made on how resources are handled
- The Fire Marshall also wanted to use historical data to identify which locations are high-risk and evaluate department performance to make decisions surrounding training, resource allocation and safety initiatives to reduce future risks
DELIVERABLES
The client wanted me to produce three reports in power BI to help answer the user stories and questions which were created and sketched by a previous client who has moved on from the project.
I reviewed the sketches and user stories, and made some slight adjustments to the design sketches to provide additional insights and make the dashboard easier to navigate. This included having a view of the top 10 and bottom 10 incident types that have occurred in the state of Georgia, instead of a sorted list of all 39 incident types, as due to the layout and size of the dashboard, the full 39 incident types would not fit on the dashboard unless a scroll feature is applied.
I also applied a filter for incident type to the map visual showing which areas in Georgia had experienced the most fire safety incidents. This will enable to the Fire Marshall to have the option to focus on specific fire incidents when comparing the frequency of fire incidents across Georgia.
For the last dashboard which shows an analysis of each fire station’s performance relative to the targets, I decided to add a line chart to show how performance for each station has changed over time. Although the previous client had included a filter for month on all the charts in this specific dashboard, I believe this creats a risk of filter blindness, hence the line chart will provide a more holistic view of how performance for each station has changed over time.
DATA
I was working with three csv files, one containing a dataset of the historical records of fire incidents that have occurred across Georgia from 2021 to 2022, one containing the mapping for incident codes to the incident type/name and a third containing the location details for each fire station.
The previous consultant working on the project had also put together a design and technical sketch of each dashboard as part of their handover notes.
What data are you working with? Live? Extracts? single/multiple tables? Structure etc. Use the data section to give an idea of the type/size and format of the data you are working with but there is no need to leave any links or file locations as people reading your log may not have access to your/your client's files.
OUTCOMES
SOFTWARE USED
The software I used for this project were Excalidraw to make plan and edit the dashboard sketches, and Power BI to create the dashboards.
WORK DONE
I produced three interactive and exploratory dashboards, as per the previous consultant’s design sketch. There was one dashboard showing the regional breakdown of fire incidents in Georgia, a second dashboard showing which day and months fire incidents are most prevalent, and a third dashboard showing which fire stations response times to incidents are longer than the target response time.
IMPACT
With the dashboards that I have created, the Fire Marshall is now able to explore and identify which regions have had the most number of fire incidents across the period of 2021 to 2022, as well as being able to explore which day and time each incident occurrence has been most prevalent. From my third dashboard, the Fire Marshall now also has a clear view of the average response times for each fire station and which ones take longer than the target.
These dashboards in turn will help the Fire Marshall to identify whether and where resources need to be distributed in order to ensure all fire stations are operating efficiently and safely.
CLIENT FEEDBACK
The Fire Marshall liked the charts that had been put together for the dashboard and also liked the interactivity that had been implemented in the dashboards. The Fire Marshall was also happy with the proposed amendments to the dashboard sketch, although implementing some of them in the dashboard were outside the scope of the project.
The Fire Marshall felt improvements could have been made by ensuring the main dashboard was fully complete, i.e. tidying up the layout of the dashboard and the filter for ‘Incident Date’.
NEXT STEPS
The next steps for the dashboards that I have created will be to complete all the visuals – specifically the KPIs for the BANs in the second and third dashboard (the dashboard analysing which day and time fire incidents are most prevalent and the dashboard that analyses the performance of all fire stations against the targets), and the additional charts that I proposed in the amendment of the dashboard sketch/plan.
I will also tidy up the layout of the dashboard, ensuring that there is enough padding on all of the elements and everything is aligned well.
On the field I created to separate ‘Incident Date and Time’, I will change the format of the date field to be able to create a date hierarchy, which will tidy up the view of the date filters by enabling specific charts to be filtered by month and day (and possibly year) instead of exact dates.
PERSONAL REFLECTIONS
SKILLS GAINED
This having been the first time I have created a dashboard in power BI, I have gained a lot of skills including but not limited to, formatting dashboard background, charts and other elements on a dashboard. I have also gained skills on how to work with spacial data in Power BI to create map visuals.
LESSSONS LEARNED
Knowing what I know now, next time I will focus on ensuring I offer at least the minimum viable product. In this project, that looks like ensuring the main dashboard is executed perfectly, i.e. ensuring that the layout, padding and format is seamless and all fields for the charts are tidy. In my instance, I would ensure that the field for ‘Incident Date’ is filterable for month and year, instead of the exact date.
CHALLENGES
A challenge I faced was working with the special data in the dataset. Power BI works with special data differently to Tableau, so this required me to play around with the different map visuals available and the various location fields to find one that was compatible.
SUCCESSES
I was able to successfully pick up work from a previous consultant in order to create a dashboard that helped the Fire Marshall solve the problems and questions he needed answers to. In addition to this, I was able to identify some areas for greater insights and exploration for the Fire Marshall within the plan that had been handed over from the previous consultant.
Below you can see the reports I produced in Power BI for the Fire Marshall.