Real World Challenges
Worth of Prizes
On the hottest weekend of the year so far, we welcomed 100 highly skilled Data Science and Project Delivery professionals to the Microsoft Reactor for Project:Hack3. Together we enjoyed a fun and engaging weekend jam-packed with data challenges, masterclasses and networking. We are proud of how our teams pushed the boundaries of project delivery to develop innovative solutions to real-world problems and so we wanted to share with you some of the output below. If you are excited by what you see, book your place now at Project:Hack4.
Output from Challenges
Following the introduction of the challenges and data sets, attendees joined teams for the duration of the weekend. With on the ground support from data scientists, PowerBI experts and subject matter experts they worked to interrogate the data, develop proof of concept and deliver a solid pitch to support it. Industry experts judged the final proposals and awarded some great prizes too! Unfortunately we had some sound problems when recording the pitches so we don’t have them all, but check out the pitches we do have as well as blogs and presentations below. They should give you a real flavour of what Project:Hack is all about!
Construction Site Image Classification
Using machine learning on photo datasets from site diaries, teams were able to identify images that contained certain materials. By quickly filtering 100,000s of photographs they were able to respond efficiently when responding to the incidents that occurred on site.
Predicting our Energy Consumption for Projects
This challenge combined two datasets:
1. Energy & Carbon usage sustainability KPIs
2. Data on who is on site before the project starts.
The Data Diggers created a dashboard to correlate the two datasets which could be used to analyse energy consumption and therefore predict energy costs more accurately prior to commencing a project.
Site Diary Quality Evaluation Tool
The D:Ream Team created a PowerApp that evaluates the ‘quality’ on diary entries and provides feedback for individuals to help them improve their entry. This ensured real time quality assurance on data entry using natural language queries.
Optimising Resource Allocation
The Concrete Snaggers performed visualisations on different datasets including a snag data set, site dates and programme data, as well as using AI functionality in PowerBI to predict snag durations. This enabled them to allocate resources to maximise efficiency.
Predicting Stalled Projects
The team analysed open data on complaints to create a model to gain insights and predict construction sites in NYC that were at risk of being stalled.
Predicting Problem Projects
Using machine learning the team were able to develop a high success rate model to predict risky projects using Gradient Boosted Trees and a Proof of Concept on cost variance.
Reducing Delivery Delays
The Tarmac Team created a dashboard to report on potential productivity savings by reducing idling of vehicle run times.
Improving Commercial Forecasting
The team developed a dashboard to report on project costs and resources, facilitating more accurate forecasting and increasing cost efficiency.
Planning US Government
The team used a U.S Government I.T Projects dataset to try and get useful indicators into the success or failure of future projects. They used techniques including natural language processing and machine learning to analyse the data, along with data visualisations in Power BI.
Videos of Masterclasses
We brought together some of the thought leaders and movers and shakers from across the industry to mentor attendees and provide expert advice on challenges. Over the course of the weekend attendees were treated to masterclasses on a range of subjects from hard core data science through to visualisations to help them up-skill. Videos of the masterclasses can be viewed below:
Join Amy for a quick-fire ‘show and tell’ of all the technologies she has found useful at previous hackathons on Microsoft Azure cloud platform. Covering technologies such as data science tooling, machine learning cloud services, business intelligence services and technologies to help you build quick proof-of concept applications to really ‘wow’ at the end of the hackathon.
This is an end-to-end tutorial of how to get started in a data based hackathon using R. How to take a problem, obtain data, produce plots and a statistical model and produce a data product, all within the one programming environment.
Find out how to build a basic data flow pipeline using Azure services. After a brief overview of the available services (Logic Apps, Data Factory, SQL server and Power BI), we will work through the process to configure an end-to-end pipeline with a few tips and tricks along the way.
Martin provides an overview of the challenge associated with project data analytics, why it is needed, how it will impact the delivery of projects and some pointers on the way ahead and how to get involved.
Once we’ve collected and analysed our data, we need to explain our insights to an audience in order to make things happen. Which types of chart should we use to communicate our data truthfully and clearly? In this masterclass, the audience are the subjects of a visualisation experiment. We’ll show you the same data but visualised in different ways – for example, as a bar, pie, line chart or scatter plot. You can vote on the visualisation which works best and explain why.
The main focus will be visual design and communicating data through choosing the right visuals, layout and colours.
Thank you to our sponsors for all their help and ongoing support.
Add Your Heading Text Here
Ticket sales from Project:Hack 3.0 raised £694 for Cancer Research UK.
The Project Data Analytics Community exists purely for the benefit of its members.