How has Nokia employed business intelligence tools? What are the effects on the decision making process?

Inbar S. would you like to share your view on this topic?

We go through a huge amount of data on a daily basis and use various business intelligence tools to structure and analyze it. In IT, for instance, we base our conclusions on the data that these tools provide and this process often triggers very productive discussion and highlights areas of improvement.

Very often categorizing the data alone will reveal some of the patterns that occur in the process - some are effective patterns that we formalize into a standard process and try to replicate across various projects/team and some are inefficient patterns which we brainstorm improvement actions for. The great thing about business intelligence in big companies is that the data samples are so large, that basically any pattern we are able to spot is significant and is worth analyzing.

thank you so much for your help

Hi Pushkar,

At Nokia we pay a lot of attention to data - we go through a large amount & various types of it daily. Data gives us insights into various patterns that occur as a result of our internal processes and serves as a basis for decisions that we make on improvements, mitigation actions, cost saving initiatives, productivity gains and so on.

In IT we use different practices to record, structure and analyze data. Those will depend on the purpose of the analysis, as well as the type and background of data. Generally, on a large scale we'll use our Enterprise Resource Planning (ERP) system where the data from around the world is recorded, categorized and stored in a network of databases. From there teams across all organization are able to use reports and analytics tools to access & display it, perform analyses, calculate KPIs, etc.

In some cases when the data is not sensitive and/or the dataset is not big we might even use third party tools like PowerBI or even Excel for quick calculations, trend analyses, testing hypotheses and patters. However, for such purposes the actual tool doesn't really matter - it's more the logic and overall approach to data analysis and decision making what counts.

Yashi, there are many examples throughout Nokia. We have an annual Artificial Intelligence conference each year where each internal organization has an opportunity to share their best practices. Highlighted areas where we have utilized AI/ML include Product Optimization, Business Process Automation /Optimization, and Customer Services.

Amritha, I like the question because it highlights the fact that there are so many applications available and I think that it's good to have a variety of options in your toolbox. Excel is still an industry leader. Everyone knows Excel. In the Business Intelligence world Tableau and Power BI are great options. In the predictive space R and Python are leaders but some drag and drop options are getting more popular. Orange, KNIME, DataRobot and RapidMiner. So there is really no single go to application that has the most impact. I think it's important to have proficiency in multiple applications because the analytics world is very diverse.

Thank you Steve for your thoughts. As a business analyst student I found Microsoft PowerBI and Tableau to be equally appealing and have a clear idea on their pros and cons. But I have a doubt on how you would find out whether Python or R is best for a specific project for predictive modelling. Do you specially use any metric in your team at Nokia to find the one is more suitable for each project?

Companies may make decisions about applications and tools based on a variety of reasons. Nokia deploys PBI because it's a good application and we are a Microsoft heavy company. Personally I like Tableau. I'm working on a project now where we have chosen Python - simply because the lead analyst prefers Python. Nokia does not have a metric to guide application selection. I would say to select the analytics tool that meets your needs. Let the project dictate your selection.

Bell Labs is leading this effort and developing advanced AI models for deployment across Nokia. Bell Labs has also led the Data Science Big Data effort by conducting regular webinars highlighting best practices in Data Science. This aligns with the overall Nokia initiatives to become a Data-Driven Organization.

Hi Rashik! I am currently not so involved in ML models or doing pure DS role activities since I work with Process Mining models but I will try to find someone who will be more capable of answering you and supporting you!

hi @Dharmendra S. , is this something you can support with?

Hello, we use both cloud hosted models like GPT 3.5 and GPT4.0 as well as on-prem hostel models like Mistral and LLaMA. The choice of model depends largely on the use case, but resource needs and security aspects also play a crucial role. I wish you the best.