Using Analytics to find Expansion Opportunities 

Transcript

Welcome to a software strategy podcast today we're talking about data science as a service. This is an SSTC software strategy case study.  It is one of the consultancies that we run.   We run data science as a service, a user experience lab, information security among others.  This was a project that we did in conjunction with Nottingham Trent University.  They run a service called Productivity Through Innovation which allows them and their clients to access European Regional Development funds.  We were lucky enough to work with Mace Sport and their very forward thinking and energetic CEO Ben Thompson.

Discovery 

We ran through our standard discovery process which really starts off just working through a given business process just seeing how we might assist in that business process;  what are the issues that the customer is facing.   Maybe it's just a procedural issue that we can just work through how the workflow might be improved.   Maybe there's some automation that might help and that we might know the market well enough to be able to recommend somebody or at least develop a short list.   It might be that we can help develop a specification or indeed run a procurement process right the way through to maybe even helping with the build-ut of that software.

The most important artifact that we develop in that in that discovery phase is a process flow.  This really defines the client engagement.  For Mace Sport we gathered,  social media data, data within Mace systems as well as government data and put it into a data lake such that we could build an operational data store, from there moving into a data warehouse.  Dashboards for management inquiry to answer executive questions and provide support through drill downs.   When we run those meetings it allows us to examine what questions should be answered within those dashboards.  

We bring together other CEOs that might have similar information needs.   We started to understand information requirements and what questions he's really interested in.  

• who is likely to convert from a spectator to a participant?

• how long will it take to convert them?

• do they need other equipment to participate in a given sport?

•  are they likely to require online or will they be candidates for face-to-face?

•  is there a given discipline in one sport that might be useful in another sport?

Conceptual Prototype

Once we have gone through what questions might be asked, we quickly develop a conceptual prototype so that we can provide something back to the clients to actually look at.   You might be able to click through it but it's not real software. It allows us to understand whether or not we have understood the issues.   It allows us to look at what the information requirements are and whether or not we have data to help answer those questions.   It allows us to like build out a specification. 

Proof of Concept

 Clients might well ask us to go further and like build it out as a proof of concept in real software. 

Analytics Model

 This allows us to understand how the analytic model will hang together; 

• what are the facts that we might need to gather together? 

• What metrics might we be able to determine by those facts?   

• What are the targets that we should have for those metrics?   

• What threshold should we have around those targets to know how well we're doing?  

• What are the dimensions around which we would want to perform an analysis?  

Transformation

It allows us to confirm that we've got the appropriate data in our data lake and to make sure that our data pipelines are actually working correctly. It allows us to go through transformation.   When we get the data from a source system it ll be in a very good transactional format but it might not be in the best format for analytic reporting so we'll go through some transformation. 

Aggregation

but analysis tends to happen at multiple different levels so we'll have to go through summarization as well so this um you're seeing here just that uh what the structured query language looks like to summarize up at different levels of summarization such that we can build out an analysis cube or a data warehouse,

Visualization

This allows us to provide what's needed for visualization so then the client can really get their hands on like real software and they can drill down uh confirm that it's working as they would like.

Professionalization

But this is just the proof of concept.   If the client wishes to move it to a professional level then we've got to go through:

• providing the specifications

•  make sure that we are writing out the technical approach

•  we've got a good qa plan in place

  •  unit testing 

  • ui testing as well as 

  • install testing 

Incorporation into Teaching

This case study was run uh in conjunction with Nottingham Trent University so it allowed us to incorporate this project into teaching.   We were fortunate enough to have Ben come and talk to students and explain why data science was helping him as an entrepreneur in creating and moving his business forward. It meant that students got the real world problem of looking through the data sources and going through data prep.  They they got to work through what the analysis questions should be.  They got to work through what an analytic design might be.  They got to work out what the best visualization might be to convey that data to management.   It allowed us to advertise the hands-on nature and the industry engagement that we had with our data science courses.

Creating Opportunities for Students

At the time this was done I was teaching full-time so it meant that we had to find partners in the local area that were experts in the platforms used to develop the application and in finding those partners it also meant that we created opportunities for graduates to support the application on an ongoing basis. 

Data Science as a Service for you

So if you think that there's something that we could help you with in the data science area,  If there's an analysis problem or a difficult business question that you wish you could get answered, then please feel free to start a conversation with us by emailing info@softwarestrategyconsulting.co.uk  We look forward to hearing from you thank you