devopsy.kinsta.cloud

Tag: data analysts

  • Logi Analytics Simplifies Dashboard Development

    Logi Analytics Simplifies Dashboard Development

    Logi Analytics launched a tool that promises to make it easier to design, build and embed interactive dashboards and data visualization within applications.

    Company CEO Steven Schneider said Logi Composer helps ease the task of creating data visualizations within the context of a DevOps workflow.

    Based on technology Logi Analytics gained via its acquisition of Zoomdata last year, Logi Composer enables DevOps teams to create interactive self-service dashboards for end users as part of a microservices architecture based on Docker containers, he said.

    To help achieve that goal, Logi Composer includes support for a range of Smart Data Connectors constructed as a set of microservices to access databases and other things.

    Schneider said as applications evolve developers increasingly are being required to embed business intelligence (BI) functionality within their application. However, he noted, most end users don’t want to move data into a separate BI application to analyze trends when they could be surfaced more easily within an application they are already using.

    The challenge has been finding a way to provide developers with a set of tools for building self-service visualizations in a modular fashion that makes them easier to build, maintain and update, he said.

    Of course, developers have been adding visualizations to applications for years. But most of those visualizations don’t let end users interrogate data—users now expect to be able to manipulate visualizations with a modern application at a much deeper level, said Schneider.

    A recent market study published by Dresner Advisory Services finds 62% of respondents consider self-service BI to be either “critical” or “very important.” A separate report published by Allied Market Research estimates the market for embedded analytics will generate revenues in excess of $60 billion by 2023.

    Data visualization is a critical aspect of every digital business transformation initiative. As organizations look to digitize more processes in the wake of the COVID-19 pandemic, the need for more robust data visualization capabilities within applications has become more acute. Most end users are unable to make sense of massive amounts of data without the aid of visualization tools.

    It’s unclear to what impact embedding BI capabilities within applications will have on standalone BI applications. There may always be analysts who want to pull data from multiple application sources to analyze trends. However, the bulk of users are likely to have their BI requirements met within an application that includes connectors to pull in external data when required. Applications that don’t include data visualization capabilities are not likely to be widely employed by users.

    Ultimately, embedded data visualization will reduce reliance on both spreadsheets and standalone BI applications. However, like most substantial changes to end user behavior, it may be a while before that cultural shift takes hold.

    In the meantime, DevOps teams should expect to see a lot more analytics functionality being embedded within applications. The challenge now is finding a way to streamline the development and deployment of those modules within the context of a larger application development project.

  • Building Your Data Science Team from Within

    Building Your Data Science Team from Within

    The global machine learning (ML) industry is increasing at a compound annual growth rate of 42% and will be worth $9 billion toward the end of 2022. Despite this figure, there are many cases of data science teams’ failure to deploy a reliable artificial intelligence (AI) product or service.

    Part of the failings and shortcomings of AI concern the ramp or process for getting there. For an AI project to succeed, it is critical to have your data fully deployed, available, structured and cleaned. You also must have some algorithms (ML/deep learning/natural language processing) ready to pull the insights and focused data, effectively putting the intelligence into AI.

    This requires a very specific skill set and talent pool.

    Empower Your Data Analysts

    With so many different technology fields colliding, machine learning is a difficult area to master. Deep reinforcement learning (DRL), natural language processing (NLP), AutoML tools, ML Ops, neural networks and model-based reinforcement learning are just some of the subjects required to master ML.

    With a shortage of talent and shortfall of data scientists, this is easier said than done. Part of this is because the demand far outweighs the supply, with Bloomberg estimating job postings for data scientists rose 75% from 2015 to 2018. Almost every organization shares this same urgent need.

    When faced with a skills gap in their teams and stiff competition in recruiting, how can organizations effectively build a data science team?

    One way is to start by looking at your data analysts. They are often swimming in disparate, federated data sources, operating on legacy databases, working with mixed data sources and using Excel to do things it was never intended to do. There is an enormous opportunity to empower data analysts to become data scientists.

    From Data Analyst to Data Scientist

    Companies looking to hire data scientists from within need to think about how to upskill, reskill or preskill their data analysts to perform the roles needed to implement AI and ML fully. I’m a believer that if you provide prescriptive and progressive curricula around the essential topics a budding data scientist needs, you can equip them with the skills and knowledge to make a difference and move them forward in the organization.

    However, it’s not enough to simply equip these employees with a curriculum; the program should have numerous twists and turns. Why? When thinking about how you can train the prototypical ML expert of the future, you must consider all the deep tech skills needed. You also want the business and social skills that are required to attain the expert moniker. A well-rounded ML architect will see the business imperatives, understand how to communicate with engineering and business, and have the tech skills to deploy ML models that benefit your business and customers.

    When organizations take roles commonly found in the workplace, such as data analysts, provide them with a sequenced path of instruction that moves them toward an aspirational role, they’ll be well-positioned to realize the full potential of AI and ML.

    — Mike Hendrickson