Notes

Points covered during the discussion

• Animating outbreaks, D3, leaflet
• Overlaying contact networks
• Including phylogenies, colouring nodes
• Static contact networks vs dynamic visualisation, first approach difficult for large data. Potential help from Rgraphviz.
• Incorporating contact times into dynamic visualisation (time directed graphs).See Eiko Yonike.
• Interactive bubble plots, visualising multi=way tables or binary matrices
• Visualise binomial GLM, GLMnet
• Visualising patterns of missing data
• Visualising heterogeneity, duration of infectiousness
• Plot heaping when reporting contacts
• Incidence time series
• Timeline of samples
• Imaging aligned sequences in DNAbin format, improve upon current options
• Plotting occurrences of same sequences in an outbreak over time
• Develop a common framework for certain graphics to convert from static to dynamic. D3 or D4 within R.
• Interaction between plot and line list, moving from plot to relevant cases on the line list easily
• Considering usefulness of static, dynamic and interactive plots for the user or developer. Gapminder/google charts
• Standardising data structures, developing static and dynamic plots separately. Turning interactive plots into static plots using webshot, converting svg into static images (chrome addon).
• Defining a standard data format, use json format and specify fields. Quick to move between R and json.
• Plot.ly , reproduce plots from other software, converts data into json for further visualisations. A potential model.
• Need to be able to produce plots without uploading data to the internet. D3 does not do this.
• D3 javascript libraries: leaflet, mapbox, sigmajs (contact networks), htmlwidgets (plug js libraries into R),crossfilter,networkD3
• Finding relevant js libraries to reproduce plots from R.