Improving the Student Experience with Advanced Analytics
Victoria University (VU) is one of the world’s top universities with more than 40,000 students, and multiple locations through Australia. The organization has been increasing its emphasis on advanced analytics, with plans for addressing academics, marketing and operational efficiency. First, the team focused on the student experience and using new data sources and analytics to identify students at risk of dropping out.
Part of VU's goal was to leverage more data sources, using not just structured data from university databases, but also unstructured data from logs generated by other student systems and facilities. Due to the data variety, volume and diversity, the analytics team knew the challenge would require more than their traditional data warehouse and SQL business intelligence (BI) tools. They were ready to try a Hadoop/Spark stack that could support all analytics, including R, Python, Java and more.
VU turned to specialist BCS Technology, for help choosing the right service, platform and deployment. The goal was a quick proof of concept that could show the potential value, then move quickly into production. BCS lead VU to an innovative solution, Cazena's SaaS Data Lake, with Cloudera's industry-leading Hadoop stack embedded.
If we had to do this on our own, it would have taken at least six months. With guidance from BCS Technology and Cazena’s SaaS Data Lake with Cloudera on Microsoft Azure, we were up in a couple of weeks. We were able to get to work on the data, without having to worry about all the other requirements of managing the infrastructure ourselves.”
The team at VU wanted the flexibility of Hadoop, but knew a DIY deployment could be long, risky and would require investments in new skills and DevOps. Using a cloud service was a viable choice, though a first for the team. However, a cloud service would ideally need to be based on Microsoft Azure, VU's IT standard. It would also need to meet strict requirements for security, compliance and student data privacy.
BCS Technology helped them identify a Cazena solution that addressed all of their challenges, in record time:
- Implement a secure platform for big data and advanced analytics quickly, without adding new DevOps or IT requirements.
- Collect, store and analyze big data from many sources; reduce cycle times from data collection to analytics and action.
- Add a foundation for analytics, focusing first on student retention by using new data sources and advanced analytics to proactively identify students at risk of dropping out.
VU partnered with BCS Technology to evaluate deployment options for big data, advanced analytics, and machine learning applications. After a fast proof of concept (POC), VU was able to move quickly into production with Cazena’s private SaaS Data Lake on Azure.
After moving into production, VU began connecting data sources, using streaming software and secure data ingest capabilities included with Cazena. Analytics teams work with Cazena via secure SaaS interfaces or connect existing BI tools to access and analyze data. The entire project was complete in weeks, with no impact to IT or additional DevOps required -- Cazena handles all operations, upgrades and platform maintenance. Everything meets regulations for compliance and privacy, with ongoing security monitoring.
By choosing a SaaS Data Lake, VU was able to focus resources on real analytics use cases months sooner than planned, delivering value faster. The results of the first project speak for themselves: VU can now identify students at risk of dropping out within a few days, based on predictive indicators. That gives VU teams a huge advantage in reaching out to students with assistance or resources to help them complete their educational plans.
Other tangible business outcomes for VU include:
- Exponentially increased analytics capabilities in weeks on Cazena SaaS Data Lake on Azure.
- Unified SaaS access to many disparate data sources into one secure SaaS Data Lake that supports all analytics and modern machine learning.
- Successfully added a foundation for ongoing analytics, with no additional resources or headcount ever required.
For us, the success of this project is not just about the technology, it’s finding the right partners that you can work with closely. Make sure you have a very tangible use case that you can measure against, and be persistent in starting small.”