Demonstration of the Gauteng Province Covid-19 Visual Analytics Tool
NA-Online
About the Event
The purpose of this session is to demonstrate the Covid-19 Visual Analytics Tool (https://gpcoronavirus.co.za/) as a management tool that can assist evidence-informed decisions and monitoring of interventions at the community level.
The tool was developed by a partnership of IBM Research (Johannesburg lab) with the Gauteng Provincial Government (GPG), Gauteng City Region Observatory (GCRO - a research institute established by the GPG and partners), University of Witwatersrand (Wits) and iThemba LABS of the NRD and University of Pretoria (UP). IBM Research developed the tool to enable spatial and temporal visual analysis of trends in the numbers of cases (total and daily), hospitalization, deaths, recoveries, tests and active cases. Some of the data were obtained from the Data Science for Social Impact (https://github.com/dsfsi/covid19za) open data repository, collated by the Data Science for Social Impact Research group at UP. Mobility data from this repository and epidemiologic hotspot models developed by Wits University and iThemba Labs also fed indirectly into the tool. Developing the tool entailed six months of intense effort and dialogue amongst partners, with a core team of two software engineers and experts in spatial data analytics and project management. A collaboration agreement of IBM Research with Wits University and further agreements with the GCRO and GPG facilitated access and sharing of data and the platform prior to its launch publicly. The partnerships remain in place and facilitate continued development of the tool as novel analytics are added to provide further insights on the evolution of the pandemic at a local-scale in Gauteng.
The tool provides data to identify development priorities at the community level and for tracking the effects on health of pandemic control measures. It provides the following data for management.
• Epidemiological data: the tool provides daily updates on the number of COVID-19 cases and deaths, enabling the visualization of their geo-spatial location from ward-level up to the provincial level for Gauteng. This was useful for monitoring the intensity and severity of the pandemic. Besides the choropleth mapping of the spatial distribution of cases, a heat map can be enabled which is a useful visual cue on the geographical clustering of cases at any point in time. The temporal filtering functionality provides users the flexibility to assess the temporal trajectory of infections during specific periods of interest.
• Social vulnerability data: the tool includes a vulnerability index that assisted in identifying areas more vulnerable to viral spread and poor COVID19 outcomes (e.g. less opportunity for social distancing, access to health care, access to social assistance). This helped to identify and assist vulnerable communities, e.g. by providing water to the communities.
• COVID-19 hotspots: Wits and iThemba LABS developed a model to predict the number of cases during waves for policy-makers to make decisions regarding resource allocation and management. A part of this effort is the hot-spots modelling, which is based on the use of Artificial Intelligence through unsupervised machine learning.
The tool is useful in a pandemic such as Covid-19, to promote more informed decision making on the social and economic challenges faced by communities, enabling government to respond appropriately and timeously. The tool can be expanded to other provinces and international metropoles. Additional risk factor metrics can also be displayed if the data is made available for the areas in question. Hotspots and predictions may also be expanded to other areas if additional data, especially geo-coded cases data can be provided to the Wits modelling team. Lastly, the architecture of the tool enables monitoring of the geographic and temporal variation of infectious diseases in general. These efforts can be expanded to provide crisis management tools and early detection algorithms.
The tool was developed by a partnership of IBM Research (Johannesburg lab) with the Gauteng Provincial Government (GPG), Gauteng City Region Observatory (GCRO - a research institute established by the GPG and partners), University of Witwatersrand (Wits) and iThemba LABS of the NRD and University of Pretoria (UP). IBM Research developed the tool to enable spatial and temporal visual analysis of trends in the numbers of cases (total and daily), hospitalization, deaths, recoveries, tests and active cases. Some of the data were obtained from the Data Science for Social Impact (https://github.com/dsfsi/covid19za) open data repository, collated by the Data Science for Social Impact Research group at UP. Mobility data from this repository and epidemiologic hotspot models developed by Wits University and iThemba Labs also fed indirectly into the tool. Developing the tool entailed six months of intense effort and dialogue amongst partners, with a core team of two software engineers and experts in spatial data analytics and project management. A collaboration agreement of IBM Research with Wits University and further agreements with the GCRO and GPG facilitated access and sharing of data and the platform prior to its launch publicly. The partnerships remain in place and facilitate continued development of the tool as novel analytics are added to provide further insights on the evolution of the pandemic at a local-scale in Gauteng.
The tool provides data to identify development priorities at the community level and for tracking the effects on health of pandemic control measures. It provides the following data for management.
• Epidemiological data: the tool provides daily updates on the number of COVID-19 cases and deaths, enabling the visualization of their geo-spatial location from ward-level up to the provincial level for Gauteng. This was useful for monitoring the intensity and severity of the pandemic. Besides the choropleth mapping of the spatial distribution of cases, a heat map can be enabled which is a useful visual cue on the geographical clustering of cases at any point in time. The temporal filtering functionality provides users the flexibility to assess the temporal trajectory of infections during specific periods of interest.
• Social vulnerability data: the tool includes a vulnerability index that assisted in identifying areas more vulnerable to viral spread and poor COVID19 outcomes (e.g. less opportunity for social distancing, access to health care, access to social assistance). This helped to identify and assist vulnerable communities, e.g. by providing water to the communities.
• COVID-19 hotspots: Wits and iThemba LABS developed a model to predict the number of cases during waves for policy-makers to make decisions regarding resource allocation and management. A part of this effort is the hot-spots modelling, which is based on the use of Artificial Intelligence through unsupervised machine learning.
The tool is useful in a pandemic such as Covid-19, to promote more informed decision making on the social and economic challenges faced by communities, enabling government to respond appropriately and timeously. The tool can be expanded to other provinces and international metropoles. Additional risk factor metrics can also be displayed if the data is made available for the areas in question. Hotspots and predictions may also be expanded to other areas if additional data, especially geo-coded cases data can be provided to the Wits modelling team. Lastly, the architecture of the tool enables monitoring of the geographic and temporal variation of infectious diseases in general. These efforts can be expanded to provide crisis management tools and early detection algorithms.
Speakers
| Name | Title | Biography |
|---|---|---|
| Mr Mduduzi Mbada | Mr | Mduduzi Mbada is the Head: Policy Research & Advisory Services Unit at the Gauteng Province Office of The Premier. Also joining the presentation are the co-developers of the system from IBM, University of Witwatersrand, University of Pretoria and Gauteng City-Region Observatory. |
| Babette Rabie | Prof | Babette Rabie is a professor in policy and public sector performance management and on the board of directors of SAMEA, the South African Monitoring and Evaluation Association. SAMEA provides a platform to share innovative practices in M&E. Closing remarks on the use of the tool will be offered. |