Hi all! After another busy week of training, DS52 were asked to put their newly attained PowerBI skills into practice in the Friday project.
PROJECT LOG
OVERVIEW
The aim of this project was to build upon the dashboard concept and design sketches produced last week and create the dashboard/s in PowerBI. I was assigned the insurance claims dataset which included information from an insurance company’s automotive division. This included policyholder information such as vehicle types, age, marital status as well as information about individual claims.
STAKEHOLDER(S)
Robbin Vernooij
Insurance Claims Specialist
BACKGROUND
The claims specialist needs to report on the latest period of claims made by their policy-holders. Along with providing important details on KPIs, they also want to highlight areas of concern such as areas of disproportionate claim or general concern.
OBJECTIVE
- Decide upon a final sketch of the dashboard that you are planning to build
- Connect to the data and build the dashboard in PowerBI
DELIVERABLES
- Published PowerBI dashboard
DATA
- Policy Claims.csv - contains ~19.1k rows of data about claim detail (4 columns: claimed policy ID, claim ID, claim number, and claim amount)
- Policy Details.csv - contains ~37.5k rows of data about policyholder details (14 columns: ID, vehicle year, vehicle colour, vehicle model, vehicle make, car use, birthdate, gender, kids driving, parent, education, household income, coverage zone, and marital status)
- Reservations.csv - contains ~98k rows of data about hotel reservations and customer details (15 columns: hotel name, reservation ID, booking date, check in date, check out date, adults, children, room type, average room rate, booking channel, days booked, stay duration, reservation status, special requests, and advanced booking)
OUTCOMES
SOFTWARE USED
Excalidraw and PowerBI
WORK DONE
- Addressed client feedback given to previous dashboard sketch and created an updated final dashboard sketch.
- Created dashboard in PowerBI with 7 charts, 4 KPIs, and 3 filters.
IMPACT
A series of charts (as well as filters and different views) have been suggested that can be later created in Tableau to produce a draft of a dashboard that would help deliver meaningful insights for the Sales and Marketing Executive.
An interactive dashboard (with filters and dynamic charts) was created in PowerBI which can later be polished and built upon to help the Insurance Claims Specialist to better understand the demographics of their customers/claims so they can advise the risk team accordingly.
CLIENT FEEDBACK
- Client feedback was generally positive regarding the charts created and the layout of the dashboard as well as the changes made to the existing sketch.
- Regarding the “with/without children” bar chart, it was advised that this would have better impact if sorted in ascending order (with 0 children at the top) to make the key insight immediately clear.
- It would be best to start with the main insights of the dashboard to make this immediately clear - this also helps the client quickly understand how their questions/needs have been met.
NEXT STEPS
To add dynamic average/reference lines that provide additional business context to the charts.
PERSONAL REFLECTIONS
SKILLS GAINED
- Use of PowerBI to quickly create/format charts and compile them into a dashboard.
- Use of PowerQuery to create calculated fields that provide additional insights for the client
- Industry/role research in a short timeframe.
- Use of Google Sheets to present research and sketches in an effective and clean manner.
LESSSONS LEARNED
- For the next presentation, whilst practising, I will aim to rehearse as if the client was in the room. I think this would help me better structure the presentation to best present the key insights (at the beginning, drawing more attention to them, etc.) in a way that emphasises how the dashboard
- When presenting a dashboard, it can be useful to go through a potential use case (select filters, talk the client through potential insights they can garner, etc.) - this shows off the features better than simply listing them out.
SUCCESSES
- All of the charts in the dashboard were successfully interactive and dynamic which allowed for more advanced filtering to be done.
- The speed of the presentation was much better than last week! More work needs to be done but this is definitely a start.