MEDLINE exploratory dashboard

In this exploratory MEDLINE dedicated dashboard, the user can visualise the aspects of the MEDLINE data more related with its own study, create new visualisation modules representing the questions monitored, and query directly the data. The dashboard that we made available feeds on that dataset through the elasticSearch index. It is composed of several interactive visualisation modules that utilises the mouse hover when interacting and provide information through mouse-over messages on several aspects of the data based on particular queries of interest (e.g. a pie chart representing the “public health” citations that talk about “childhood obesity” during a selected period of time; or a bar chart showing different concepts included in the articles related to “mental health” in Finnish scientific journals).

This dedicated MEDLINE dashboard serves the less technical user to explore the data available (over a subset of the data generated by a topic of interest), other options are available that permit more control of the data by the data scientists at a more operational level. These are: (i) the management dashboard, where the technical user can perform the appropriate subsampling based on the topics of interest and the optional advanced options over the available data features; (ii) the visual modules creator permitting the less technical user to easily create new interactive visualisation modules; (iii) the live dashboard, that can be set up through iframe as a live window of the decision-maker’s workflow, enabling to monitor the status of the KPIs represented at each visualisation module.

This dashboard is composed by:

• a series of dashboards that can be used by decision-makers as monitoring dashboards each of which representing one study/topic (e.g., childhood obesity)
• a set of visualisation modules that can be used to construct the dashboards in (I) and each of which represents a question to be monitored, with the tool to create new modules in predefined formats
• a dashboard with the visualisation of the main aspects of the raw data, and a query box that can be used to directly query over the dataset

To deploy and run this service the user needs to install elasticSearch and Kibana on the assigned server and download the MEDLINE dataset using an FTP connection. Then the data must be clean, preprocessed and loaded to elasticSearch with the scripts available in this repository, in order to build the index that is referenced by Kibana.

• Installing elasticSearch
• Installing Kibana
• Downloading the MEDLINE/PubMed dataset
• Cleaning, preprocessing and loading the data

To utilize this dedicated dashboard, the appropriate workflow is as follows:

1. The user will be presented with the monitoring dashboard with options on the left side bar
2. The user can start by exploring the available visualisation with mouse over
3. The user can then edit the dashboard adding other available visualisations by choosing “edit” in the top bar
4. The user can also share the live dashboard over iframe (to integrate, eg., a website or a monitoring tool) by choosing “share” in the top bar
5. To edit an existing visualisation module the user must click on Visualise on the left side bar and then the visualisation name
6. To build new visualisation module the user must click on Visualise on the left side bar and then the red plus button
7. To query directly the dataset, the user choses Discover on the left side bar and then uses the search box (the accepted query language is Lucene)

In the following we represent the workflow on the creation of the monitoring dashboard per topic:

• Exploring the raw data
• Creating-a-visualisation-module
• Building and sharing a monitoring dashboard

For more information on this service please contact [email protected].