Skip to main content
Search form
  • 00:00

    [MUSIC PLAYING][DATA VISUALIZATION, case study][Visualizing Data About UK Museums--Bar Charts,Line Charts and Heat Maps]

  • 00:16

    ALEXANDRA POULOVASSILIS: Hello.My name is Alexandra Poulovassilis.[Alexandra Poulovassilis, Professorof Computer Science, Birkbeck College, University of London]I'm a Professor of Computer Science at Birkbeck,which is part of the University of London.I undertake fundamental research in computer science workingwith other computer scientists.But I also enjoy working with scholars from other disciplinesto investigate their research questions, particularly

  • 00:38

    ALEXANDRA POULOVASSILIS [continued]: as arising in the management, accessing,and visualization of data, which ismy area of research interest.[The Mapping Museum Project]Over the past four years, I've been workingon the Mapping Museums project.

  • 00:58

    ALEXANDRA POULOVASSILIS [continued]: This is an interdisciplinary projectfunded by the Arts and Humanities Research Council,involving scholars from multiple disciplines, museum studies,geographical information systems, and computer science.It is led by Fiona Candlin, who isProfessor of Museology at Birkbeck's History of Art

  • 01:19

    ALEXANDRA POULOVASSILIS [continued]: Department.[Fiona Candlin] The aim of the Mapping Museums projectwas to research the emergence and development of the UKmuseum sector since the 1960s to the present day.We were particularly interested in the wave of small museumsthat opened since the 1970s, about which very little was

  • 01:39

    ALEXANDRA POULOVASSILIS [continued]: known.In order to undertake this investigation,we needed to collect data from over 15 primary datasources to codify and integrate that datainto a single database that could then be used to supportthe research of the project.We collected information about over 4,000 museums

  • 02:02

    ALEXANDRA POULOVASSILIS [continued]: to store in our database, which is doublethat of any other previous study or data set.The research questions that the databaseneeded to support included looking at periods wheremuseums opened a lot, looking at variations in museum openings

  • 02:23

    ALEXANDRA POULOVASSILIS [continued]: and closings, according to attributessuch as museums governance--for example, are they a national or local authorityor an independent--their location, their size, or their subject matter,which is their thematic focus.Another research question relatedto looking at distributions of museums

  • 02:45

    ALEXANDRA POULOVASSILIS [continued]: in the different countries of the UKand in the regions of the countries.[Uncovering the Researchers' Needs]In order to use the database to undertake the research,we needed to develop a web application for the scholarsto use.

  • 03:07

    ALEXANDRA POULOVASSILIS [continued]: The web application comprises three main facilities--facilities for browsing through the data; facilitiesfor searching the data using the key attributessuch as location, subject matter, governance, size,and also year opened and year closed;and facilities for visualizing the data.

  • 03:29

    ALEXANDRA POULOVASSILIS [continued]: Looking particularly at the visualization facilities,before we began to design these, weneeded to understand the detailed questionsthat the scholars needed the visualizations to support.So we asked them to list the research questions.And in doing so, three different types of questions emerged.

  • 03:51

    ALEXANDRA POULOVASSILIS [continued]: The first type of question relatedto museums that were open at a given time or thathad opened up to a given time.And the scholars were interested both in total numbers,but also breaking the numbers downinto categories such as governance, subject matter,location, and size.

  • 04:12

    ALEXANDRA POULOVASSILIS [continued]: And they were interested also in similar informationabout museum closures.[Type 1- Finding out the number of museumsthat are open at a given time, or thatopened up to a given time]The second type of question related to seeing museumopenings from the 1960s to the present dayand also number of museums open from the 1960sto the present day, so as to be able to spot patterns

  • 04:33

    ALEXANDRA POULOVASSILIS [continued]: in the trends of openings.And similarly, for closings from 1960 to the present day.And again, the domain experts wereinterested both in total numbers and in breakingdown the numbers according to the major attributes,such as location, subject matter, governance, and size.[Type 2- Seeing the pattern of museums open

  • 04:54

    ALEXANDRA POULOVASSILIS [continued]: and museum openings over all the years from 1960to the present day]And the third type of research questionrelated to cross tabulating numbers of museums thatare open at a given time, according to two categories.Two of the categories of location, size, subject matter,governance, and some others.And again, being able to drill down into

  • 05:17

    ALEXANDRA POULOVASSILIS [continued]: subcategories of these categories.[Type 3- Cross tabulating the numberof open museums at a given time accordingto pairs of categories]Having understood the different types of research questions,we then proceeded to present the domain expertswith a palette of possible choices of visual chartsthat might support each of these types of question.We then all worked together to decide which kind of chart

  • 05:39

    ALEXANDRA POULOVASSILIS [continued]: was best suited to which type of question.In making these decisions, we were very much drivenby the needs of the scholars to have the visualizations beingas simple and easy to use as possiblewhile still being able to meet their research aims.[The Mapping Museum Visualizations]

  • 06:04

    ALEXANDRA POULOVASSILIS [continued]: For type one questions, the choice we came up withwas a bar chart. [bar chart] Thisallows us to use the x-axis for different valuesof the selected attribute, and the y-axisto show the number of museums.We also added a temporal slider along the bottomto be able to change the year.

  • 06:27

    ALEXANDRA POULOVASSILIS [continued]: Bar charts are very useful for easily comparing visuallythe numeric values of different categories.In our case, the museum attributes governance, subjectmatter, size, and location.Adding a slider to a bar chart allowsyou to vary a second attribute of the data.In our case, the time.For example, in the Mapping Museums visualization facility,

  • 06:51

    ALEXANDRA POULOVASSILIS [continued]: selecting number of museums, open at a given time, all,we see the total number of museums open in a given year.The default year is the present year.And we can use the slider beneath the chartto change the year being viewed.Selecting now subject matter, we see the number of museums

  • 07:12

    ALEXANDRA POULOVASSILIS [continued]: open in a given year broken down by subject matter.Again, we can use the slider beneath the chartto change the year being viewed.Selecting buildings, we see the numberof museums relating to buildings open in a given year, brokendown by different types of buildings.

  • 07:32

    ALEXANDRA POULOVASSILIS [continued]: Finally, selecting houses, we seethe number of museums relating to housesopen in a given year, broken downby different types of houses.As another example, selecting numberof museums that opened up to a given time,governance, we see the number of museumsopen up to a given year, broken down by governance.

  • 07:54

    ALEXANDRA POULOVASSILIS [continued]: We can again use the slider beneath the chartto change the year being viewed.For type two questions, seeing the patterns of museumsthat are open or that opened over time, wechose line graphs. [line graphs] Theseallow easily to view the trend in a numeric property

  • 08:16

    ALEXANDRA POULOVASSILIS [continued]: over time.And also, it allows us to use multiple linesto see the trends for different values of the selectedattribute.Line graphs are an intuitive way of showing the trendof a numeric value over time.Plotting multiple lines on the same graphallows side-by-side comparison of several trends.

  • 08:38

    ALEXANDRA POULOVASSILIS [continued]: In our case, we use multiple linesto show the trends for different values of the museum attributesgovernance, subject matter, size, and location.For example, in the visualizations facility,selecting number of museums open over time, all,we see a line graph of the number of museums that

  • 08:59

    ALEXANDRA POULOVASSILIS [continued]: are open each year since 1960.Selecting subject matter, we see the numbers broken downby subject matter, with one line on the graphfor each subcategory of subject matter.As another example, selecting numberof museums, openings over time, all, wesee a line graph of the number of museums that

  • 09:21

    ALEXANDRA POULOVASSILIS [continued]: opened in each year since 1960.Selecting governance, we see the number of openings broken downby governance, with one line on the graphfor each different type of governance.Selecting independent, we see the numberof openings of independent museums broken down further

  • 09:42

    ALEXANDRA POULOVASSILIS [continued]: according to the different types of independent museums.For type three questions, which require cross tabulating valuesfor pairs of categories, [cross tabulating] we chosea heat map. [heat map] A heat map allows you to place in eachcell a numeric value which depends on the valuesof the two chosen attributes.

  • 10:04

    ALEXANDRA POULOVASSILIS [continued]: You can also have an additional slider on the bottomto vary a third attribute.Heat maps are a natural way of visualizinga quantity whose size is dependent on two dimensions.The cells of a heat map are colorcoded so as to give a visual cue about the size of the quantity.

  • 10:24

    ALEXANDRA POULOVASSILIS [continued]: In our case, we used a different hueof green to color code the cells,as well as giving an actual number within each cell.Adding a slider to a heat map allowsyou to vary a third attribute of the data,in our case, the time.For example, in Visualize, selecting x and y on the table

  • 10:47

    ALEXANDRA POULOVASSILIS [continued]: exposes the choices available for the two categories.These include governance, subject matter, size, location,and some others.Selecting location on the x and size on the y,we see the number of museums of different sizes locatedin each part of the United Kingdom.The slider beneath the chart can be

  • 11:09

    ALEXANDRA POULOVASSILIS [continued]: used to change the year being viewed.Selecting now Wales, still on the x,we see the number of museums of different sizes locatedin each region of Wales.Overall, our design choices for the visualizationswere driven by the needs of the domain experts, whowanted the visualizations to be as simple and easy to use

  • 11:31

    ALEXANDRA POULOVASSILIS [continued]: as possible while still being able to meettheir needs for their research investigations.We did try out some other charts as well,such as pie charts and tree maps.[pie chart, tree map] But in the end,the domain experts opted for the simplicity and visual clarityof the bar charts, line graphs, and heat maps that we selected.

  • 11:54

    ALEXANDRA POULOVASSILIS [continued]: We were also mindful that there wouldbe a wider audience for our web applicationand database beyond the project's own domain expertsfrom the UK museum sector.And we wanted the visualizations to beeasy and simple to use for them too,without requiring any specialist knowledge

  • 12:15

    ALEXANDRA POULOVASSILIS [continued]: or technical training.We did carry out a series of user evaluation exerciseswith stakeholders from this communityto test out the visualizations and the other parts of the webapplication.And we received positive feedback from this community,too.[Technical Implementation]

  • 12:38

    ALEXANDRA POULOVASSILIS [continued]: In terms of the implementation of the visualizations,the web application within which the visualizationsare implemented has a classical three-tier architecture.There is client software that runs in a user's web browser.There is server software, which runs

  • 12:58

    ALEXANDRA POULOVASSILIS [continued]: in a web server at Birkbeck, and thisconnects to the database server, also at Birkbeck,hosting the database.The client software is written using JavaScript,and the server software is written using Python.The visualizations are implemented using the Bokeh

  • 13:20

    ALEXANDRA POULOVASSILIS [continued]: Visualization Library.[Bokeh Visualization Library][Conclusion]Thank you very much for watching.I hope it has been interesting for youto see how we work together as a cross-disciplinary teambetween computer science and museum studiesand geographical information systems

  • 13:42

    ALEXANDRA POULOVASSILIS [continued]: to build a system and visualizations whichdoes meet the needs of the domain experts, able to be usedto undertake their research, and also brings out the appealand visual clarity of the selected simple visualizationsthat we did use.

  • 14:03

    ALEXANDRA POULOVASSILIS [continued]: [Further Reading,,http://museweb.d,http://blogs.bbk 0/interdisciplinarity/][MUSIC PLAYING]

Video Info

Publisher: SAGE Publications, Ltd.

Publication Year: 2021

Video Type:Case Study

Methods: Data visualization, Data management, Databases, Geographic information systems

Keywords: bar charts; charts (data visualization); communication aids; data collection; data visualisation; graphs (data visualization); interdisciplinary studies; maps; museums; research design; research questions; Software; web browsers ... Show More

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:



Alexandra Poulovassilis, a professor of computer science at Birkbeck College, University of London, discusses the Mapping Museum project and her research visualizing data about UK museums, including the needs of researchers and technical implications of data visualization.

Looks like you do not have access to this content.

Visualizing Data About UK Museums: Bar Charts, Line Charts and Heat Maps

Alexandra Poulovassilis, a professor of computer science at Birkbeck College, University of London, discusses the Mapping Museum project and her research visualizing data about UK museums, including the needs of researchers and technical implications of data visualization.

Copy and paste the following HTML into your website