You are currently browsing the category archive for the ‘Research Study’ category.

Ok, so I know I said I wanted a first draft up last Friday…but it seems I was a bit too optimistic in setting that milestone.  I am now targeting next Friday (not this Friday, but the Friday after…I’m never sure how that works).  In a meeting with my advisor this morning I was saying how I’m not sure how to judge my progress. I feel like I am getting work done, and making progress, but as I’m not sure exactly how much work is ahead of me it’s hard to say if I’m on schedule or not.  He replied with asking me how many pages I have right now that I could show him.  To which I said zero (eek!), they are all notes and point form, not really in any readable format.  Ah, nothing like a talk with your advisor to give you a reality check.  I’d say out of the 5 day week I am devoting to school, I feel good about the productivity of maybe 3.  I’m going to have to do better than that.

Last week I reviewed the results again…and then reviewed the results AGAIN…and again.  Needless to say, I need a break from my results and discussion.  I fine-tuned the classification scheme I presented earlier, and by fine-tuned I mean obsessed over.  I found it pretty tough because there is no clear way to say this is how you describe a map, there are so many factors and exceptions that play into it so I can interpret it many different ways.  I can be pretty indecisive so I need to just pick a structure and stick with it and ensure I cover all of the decisions I made in my writing.

Up until last week I had only been looking at keywords and tracking how often they were being used in the descriptions.  Now I have taken a look at the context of the keywords which allowed me to see how the categories were used in order to create a description.  I guess if you take only one thing away from my research, this should be it (albeit in it’s first draft form)…

A description of a thematic map relays context through stating:

  • Query data, for a general understanding of what the map is representing
  • Jurisdiction, for a specific area the map is covering
  • Location, to give real-life names to the area of jurisdiction

The points shown in the thematic layer are described in relation to the various topographical features and landmarks through comparing:

  • Size
  • Distance
  • Direction
  • Placement

These descriptions are enriched through the use of descriptive words of the various layers and their features, which relay:

  • Shape
  • Size
  • Colour
  • Direction
  • Quantity

There were also two noticeably dominant techniques when describing the map. To choose a specific target (ie. river, lake, city center) and then describe everything on the map in relation to the target. If there was no obvious target available, participants chose to divide up the map (into quadrants, top-down, left-right) and progressively describe the maps in smaller pieces.

Here is an example that I have pieced together from the results to illustrate what I have stated above:

to be determined

This map is showing locations of facilities which reported pollutant releases in Canada in 2008.  There is a town named Springfield, which is based along the east bank of a river.  The river runs alongside the town from north to south.  Approximately half way down, the river is joined by a smaller river that arches from the left.  At the point where the two rivers join is a railway track.  There are 3 facilities shown, the first is at the terminus of the railway track.  The second point is located at the south end of the town and to the east of the first point.  There is a small park near the center of town and bordering the east side of the river, as well as an orange symbol over the river about X km north of the park.  The third point is due east of these two features, and is just outside the north-east side of the town.


Query Data =  “locations of facilities which reported pollutant releases in Canada in 2008”

Jurisdiction = “town”

Location = “Springfield”


Size = “smaller river”

Distance = “orange symbol over the river about X km north of the park”

Direction =   “The second point is located at the south end of the town and to the east of the first point.”

Placement = “There is a small park near the center of town and  bordering the east side of the river”

Descriptive Words:

Shape = “that arches from the left”

Size = “There is a small park”

Colour = “orange symbol”

Direction = “The river runs alongside the town from north to south

Quantity = “due east of these two features”


Ok so you saw my rough wordles from the map descriptions, and the classification system that I drew up as a result of these.  Now I want to add the backgrounds of the participants that described the maps to see how the rankings of the categories are influenced.  So far I have divided the participants occupations into:

  1. Geography
  2. Information Technology
  3. Academia
  4. Other

I also classified their knowledge of Cartography, Web-Mapping, Web Accessibility and Visual Impairment into:

  1. Little to None
  2. General
  3. Expert

I made some charts using swivel over the weekend and I have been procrastinating putting them up because I’m not very happy with them, or swivel, as I feel like I spent all sunday formatting data…but for now it will have to do.  Plus they aren’t accessible!  I will re-post the participant data as soon as I have the patience to reformat the data into html tables…I promise.  But for now here are the graphs in swivel *grumble*.

The ranking of the categories used in the map descriptions from most popular(1) to least popular(12) are:

  1. Direction
  2. Geography
  3. Qualitative
  4. Shapes
  5. Landmarks
  6. Query Data
  7. Thematic
  8. Size
  9. Jurisdiction
  10. Colour
  11. Distance
  12. Location

The study provided 8 different maps that were randomly presented to the participants to describe.  The maps varied in two factors in order to measure any difference in map descriptions.  In general, these two factors were Jurisdiction, some showed a rural setting, others showed an urban setting; and Extent, some showed a zoomed in extent showing more roads etc. others were zoomed out showing entire provinces.  The following tables show how the category ranking changed depending on these factors (sorry for the horrible html tables).  For both factors the top two categories do not change, Direction and Geography remain the most popular.

Ranking of Categories depending on Map Jurisdiction
Rank Overall URBAN RURAL
1 Direction Direction Direction
2 Geography Geography Geography
3 Qualitative Qualitative Landmarks
4 Shapes Shapes Qualitative
5 Landmarks Query Data Shapes
6 Query Data Landmarks Thematic
7 Thematic Jurisdiction Query Data
8 Size Thematic Size
9 Jurisdiction Size Jurisdiction
10 Colour Colour Colour
11 Distance Distance Distance
12 Location Location Location


Ranking of Categories depending on Map Extent Level
1 Direction Direction Direction
2 Geography Geography Geography
3 Qualitative Qualitative Shapes
4 Shapes Landmarks Qualitative
5 Landmarks Shapes Landmarks
6 Query Data Thematic Query Data
7 Thematic Query Data Thematic
8 Size Size Size
9 Jurisdiction Jurisdiction Jurisdiction
10 Colour Colour Colour
11 Distance Distance Location
12 Location Location Distance

I don’t know if these are interesting to anyone else, but I thought I would put them up just in case 🙂 Each wordle shows the dominating words used within each category.








Query Data





Too many wordles? Ok ok  I’ll stop….

I have been going through my data over the last month, and while I’m not where I want to be just yet, I thought I would blog as a bit of a checkpoint…get something out there so I can review and ensure I’m on the right track.  I do not have any numbers just yet, at this point it’s solely a qualitative review.

Final Count

Participants: 122

Decriptions: 285

The approach I am taking in my analysis is Grounded Theory.  After a first pass through the map descriptions I’ve classified the words used into the following categories:

  • Jurisdiction
    • Extent: province, city, town, etc.
    • Community: neighbourhood, industrial, residential etc.
  • Geography (any reference to the base map)
    • Landmass: inlet, island, shoreline etc.
    • Water Body: lake, river, estuary etc.
    • Topography: mountain, hill, vegetation etc.
    • Road: road, street, highway etc.
  • Direction
    • Relative: left, right, up, down etc.
    • Compass: north, south etc.
    • Miscellaneous: running, tip, flowing, crossing etc.
  • Thematic
    • General: points, dot etc.
    • Specific: polluting sites, facilities etc.
  • Colour
    • Blue, orange etc.
  • Shapes
    • Literal: circle, square, diamond etc.
    • Implied: jagged, smile-shaped, meandering etc.
  • Size
    • Small, large, tiny etc.
  • Query Data
    • Facilities Reporting Pollutant Releases, 2008, Substances etc.
  • Landmarks
    • Airports, parks, ferries etc
  • Distance
    • Km, cm, inch etc
  • Quantitative
    • Specific: number
    • General: many, few etc.
  • Location (Specific)
    • Montreal, St. John, Charlottetown etc.

Some of these categories may seem related or overlapping, yet as analysis is continued they will be further refined and separated out, which will result in a clearer definition of each.  In order to fit the words properly into each category, the context was taken into account, which means one word could have been coded into more than one category.

The important part here is not how many words were found under each category, as that is basically just showing vocabulary size.  It is important to understand how often each category is used amongst the descriptions, currently Jurisdiction, Direction, and Shape are used in almost every description.  Yet this is not a very stable result as various factors still need to be taken into account, which will be summarized at the end of this post.

So what is the theory behind these categories?  The use and consideration of these concepts is imperative in order to create a meaningful map description.  From the data generated by this study, there was no “silver bullet” description.  These categories will be used as the concepts which influence the design of a meaningful map description.  They take into account both explicit (query data, distance) factors that can be extracted from the map data and also implicit (shape, direction) factors that are traditionally only available through the visual interpretation of a map.

Test descriptions will be created and then reviewed within the visually impaired community in order to assess the relevance of the categories.  It will allow us to better refine and rank the categories, as some may be more useful to blind people (i.e. direction) than others (i.e. colour).

Also, once we then take into account the purpose (why the user is coming to view the map in the first place) of the description, we can then tailor the description used in order to better suit the user.  This will result in modified categories depending on the purpose.  *This part is most likely out of scope for my research.

Factors that still need to be taken into account in order to further refine the categories are how the dominating categories change as extents change.  When dealing with a map of a city, are certain categories more imperative as opposed to those used when describing a province-wide map?  Also, these categories are based on single words and their context.  There is something to be found in sentence structure in the map descriptions.  Some take a very factual approach, which could be comparable to the results you would receive from a text-based search.  Others are quite poetic, using words that seem to convey the movement and dare it be said – the beauty of maps.

The second part of the analysis deals with the participant’s background.  How are the categories in each description influenced by the background of the participant?  If you took the study, you will remember that each participant was asked to detail their familiarity in various topics that were thought to influence the way they chose to describe the map. This will possibly affect the categories used, depending on the participant’s area of expertise.  It may also affect the approach they used in their sentence structure.  This is the second part of my analysis that has yet to begin…it’s a long wknd right? Stay tuned for a follow-up blog post…

My advisor-motivator-friend Greg Wilson has recently turned his attention full-time to working on his Software Carpentry course.  He is blogging about his work on the course here and posts updates here.

I am writing in hope that this post will extend to the far corners of my office building in Downsview and reach the dark and dusty cubicles of the scientists at Environment Canada, who are busy coding away in fortran.  This could be the answer to your prayers.

If you want anymore information don’t hesitate to ask me or contact Greg…he’s a pretty pleasant guy to chat with 😉

Take a look at the mission statement for a full description of the course.

Just a little visual delight to tide you over until I have my results in.  The data is rudimentary, but still fun to see.

A data visualization of words used to describe a map.

Map Description Wordle

Background for people reading this blog for the first time: I have conducted a research study on how people describe a map.  The above is a data visualization listing all of the words used in the descriptions.  If you are interested in the study please visit

The good news is over 100 people participated in my study, which resulted in over 300 map descriptions, and they are still accumulating!

The bad news is over 100 people participated in my study, which resulted in over 300 map descriptions, and now I have to sort and analyze the data.

That’s what I’m doing now, and although it is a daunting task I am finding it fascinating.  The design of this study entailed alot of planning and discussion (thanks Greg and Jon).  To finally see the results of something you have worked on for so long and feel connected to is pretty rewarding.  There’s alot of data to get through but it’s been fun so far (although I have just started).  I know not everyone else is as invested in my topic, but I find myself wanting to tell somebody whenever I come across something interesting, or validating, or curious, or…well I just want to tell people about everything 🙂  I am looking forward to generating my results, so I can actually do just that, and see what the web accessibility community thinks.

So this post was just an update really, to whoever follows, to let you know where I’m at.  I’ve given myself about a month to establish results.

Data from Star Trek TNG with a cat and computer.  Data is saying "No Spot, you many not 'has cheezburger.'  Not until you are able to ask in a manner that is grammatically correct and lacking typos."


My study has been released for a couple of weeks now and I am happy to report that I am reaching a large population and generating some great data.

I wanted to call to attention the fact that the study is asking people to translate their visual understanding of maps into text.  If you have any sort of  visual impairment that makes you unable to view the maps in the study but are interested in contributing I would love to hear from you.

I would really appreciate any help that will guide me through this data once it has all been accumulated.

Please send me an email at

Thank you!

Subjects are needed to take part in a study concerning Web-Mapping Accessibility.  Participants will be asked questions concerning their interpretations of maps and cartography in order to help improve the accessibility of web-mapping applications, specifically to help the visually impaired.  Time needed for the study is flexible and requires no prior knowledge.

Please visit:

Thank you!