COVID Notebooks Aims to Speed Predictive Models
IBM’s new open source toolkit with AI extensions to the Jupyter notebooks data science development platform is being extended to a COVID notebooks platform designed to help analyze real-time data about the pandemic.
The company’s Center for Open-Source Data and AI Technologies developed the COVID notebooks toolkit that among other things addresses data quality issues related to coronavirus analytics. Along with compiling “authoritative” data on the pandemic, the IBM unit said it “clean[ed] up the most serious data-quality problems.”
“Policy makers are asking questions including: What stories can we tell in the aggregate? Are there patterns we see across the country? What regions or demographics are getting affected the most by the pandemic?” the company said in a blog post.
Given that underlying data about the pandemic changes daily, COVID notebooks allows data scientists to concentrate on building models rather than data cleaning. The tool allows frequent updates of results on analysts’ notebooks.
The open source pipelines Elyra and its visual editor along with Kubeflow can be used to update results with fresh data.
Data sources include the Johns Hopkins University COVID-19 Data Repository, which includes county-level information on the pandemic. The Johns Hopkins data set is widely used to develop predictive models for national and state forecasts of deaths attributed to the novel coronavirus.
Other data sources include agencies like New York City’s Department of Health and Mental Hygiene, which includes borough-level data on the early epicenter of the pandemic.
In one scenario, IBM said data could be analyzed to detect correlations between poverty and infection rates. “Open source developers and data scientists can easily build on these tools to extend the analysis to their individual use cases,” notebook developers added.
In May, IBM announced an extension of its Elyra AI Toolkit to the industry standard JupyterLab user interface with the goal of simplifying development of AI and other data science models. The initial release included a visual editor for building AI pipelines along with the ability to run interactive notebooks as batch jobs. Other features include Python script execution and a “hybrid runtime” capability based on Jupyter notebooks’ enterprise gateway.
The COVID-19 toolkit incorporates Jupyter notebooks and Python data science libraries, including Panda. Panda data frames were used for cleaning and data analysis. IBM said it is extending Pandas for natural language processing applications.
Meanwhile, a graphical workflow editor built as part of the Elyra project ties the COVID-19 notebooks into workflows for running daily updates. Those data are collated into an appropriate format for easier analysis with tools like Pandas, IBM said.