CTI Data Lab_
Have you identified any opportunity? Would you like to test it without investments in people and technology?
Regardless of your company’s demand and maturity in this cycle, our data science team can help you build specific custom functional prototypes using a coherent methodology that identifies issues, opportunities, qualitative gains, and tangible expected results. If you have only one problem, we’ll fix it for you. To enable an efficient prototype building process, CTI has a laboratory with statistical, optimization, simulation, data mining, machine learning, and artificial intelligence technologies. With our lab, we can combine different techniques, systems, and programming languages to fulfill the specific needs of every designed prototype.
- Statistical platforms: although big data have become an organizational hype, most organizational issues still rely on low or medium volumes of structured data. In this context, the effective application of statistical concepts is extremely important. To address these problems we use R, Python and the IBM SPSS platform in a consistent and interconnected set of statistical functionalities, from efficient data processing (ETL) to statistical inference.
- Development Platform: even in the specific context of data science, the development of small applications is often seen, for example mobile and web applications. In these cases, we use tools such as Python, Node.js, and IBM DevOps development platform, combining flexibility and efficiency in concept testing.
- Artificial Intelligence Platforms: we don’t want to reinvent the wheel. We have a large library of ready high-capacity tools (APIs), mainly based on IBM Watson and Google Cloud. With these APIs we can quickly build intelligent, scalable solutions that use chatbots, voice recognition, image or video, among other resources.
- Execution Platform: to ensure solution scalability, we primarily use the IBM Cloud, based on the Watson Studio platform. However, in specific cases, we may migrate or develop our prototypes on other cloud environments, such as AWS Cloud and Google Cloud.
- Simulation Platforms: with different applications, this simulation offers a wide variety of specific techniques for each problem, including discrete event simulation solutions (SIMUL8 and ARENA) and system dynamics simulation (Python and Vensim). For Monte Carlo simulation cases, we develop specific applications using Python.
- Optimization Platform: in terms of optimization, we use the best tool – CPLEX – the basic platform for IBM Ilog and IBM Decision Optimization. With this tool, we are able to address large linear, nonlinear, integer or mixed optimization problems, combining these optimization capabilities with intelligent solutions through an integration with Python.
- Visualization Platform: the visualization theme covers many aspects. In our lab, we use IBM Cognos BI due to its simplicity and integration with the platform as a whole, in addition to its specific integration with Watson Studio. However, in specific cases, we may use customer platforms like Microsoft Power BI, Qlik, or Tableau.
- Database: different data science applications have different data storage and handling needs. For this reason, we use IBM DB2 as a structured database, IBM Object Storage for unstructured storage, Hortonworks as a Hadoop platform, and Couch DB or Mongo DB for specific low latency applications. In addition, we use IBM Planning Analytics TM1, which can immediately integrate data science solutions with the organization’s integrated planning.
By combining our technological laboratory with our agile methodology to build functional prototypes, you will be able to test your idea with the help of an expert team, in a fast, easy, and efficient manner. Headed by partners with extensive business experience, our technical team has:
Responsible for combining engineering and data science tasks with a proper technology to meet cost and performance requirements; they also prepare the solution to scale.
Understanding complex relationships requires proper visualization techniques. Our BI experts ensure not only the best way to visualize automatically generated insights but also the performance of solution reports.
Some cases generated in our laboratory:
Intelligent Computational Interaction
Problems: increase results or reduce interaction costs between organization (for example, sales force) and consumers. Acquire consumer or market data with minimal friction.
Solution: development of intelligent ChatBots or VoiceBots that can react to consumer needs in natural language.
Problems: automatically segmentation of consumers, objects, points of sale based on their behavior, with structured or unstructured data. Classify every consumer’s potential or credit risk.
Solution: machine learning application for automated online segmentation and/or classification of consumers as they interact with the organization.
Automated Suggestion of Optimal Portfolio
Problems: based on revealed consumer tastes acquired for example via intelligent computational interaction, analyze and optimize the product portfolio to maximize sales. Automate portfolio segmentation at the PoS according to consumer behavior.
Solution: machine learning solution to suggest an optimal portfolio based on the functional attributes of products, restrictions and identified/revealed desire of the consumer.
Have you identified any opportunity? Would you like to test it? Talk to us and use our laboratory!