https://www.maptiler.com/maps/#784b0675-dce5-4a4f-b16c-0d8e06442314/

• Follow github docs
• Update GitHub Actions workflow: change this line to remove the command line arguments -- --base "/itsb/":
  

  itsb/.github/workflows/gh-pages.yml

  Line 39 in 58050c0

  run: npm run build -- --base "/itsb/"

The line in the workflow was necessary because the current site is served from https://performant-software.github.io/itsb/ instead of the root https://performant-software.github.io/. This assumes the site is meant to be served from the root sameboats.org/.

When successfully deployed:

• Update the readme to use the new URL:
  

  itsb/README.md

  Line 7 in 58050c0

  For more information and credits, visit the current deployment of [_In The Same Boats_](https://performant-software.github.io/itsb/).

I looked at what happens when selecting a place in the original app design. The original app has a popup displaying the list of waypoints. (It seems to be buggy and doesn't always work. Also, it tends to get unwieldy for more than one or two waypoints.) At the moment, we only have a hover tooltip with the placename. Question is if we want to add something in our app, too?

Another, only half-related issue: when clicking on a place in the Trajectories view, this place gets globally selected, which then also causes it to be the current selection when switching to Intersections view. This seems confusing, especially because there's no visual feedback when clicking a place in the Trajectories view. I vote for making selections "local", i.e. clicks in the Trajectories view don't affect selection in the Intersections view. (Disclaimer: I'm the one who introduced this behavior in the first place, i.e. take the blame ;-)

Update: I found the original version now, so much of this is now clarified thx to issue #9 (I still leave it here for the record).

I've been wondering about what detail information to show for an intersection place. I think this will also influence the exact API / datastructure we should be aiming for. Some thoughts...

About place markers:

• If there are more than two itineraries selected, should the map show only places where all selected authors are in the same place? Or all places where more than one author was at the same place at the same time?
• Right now, lower likelihood results in larger dots with higher opacity. Should higher likelihood produce larger dots instead?

About a hover or popup component:
Currently, there's only a debug hover tooltip. As soon as more than two authors are involved, information can become a bit complex. Would some sort of calendar heatmap (like Github's activity chart) be an option? Or is this too much detail for this project?

I'm just going over the differences between our implementation and the original app. One thing that IMO looked nice in the original design was the author list. It did not have checkboxes, but instead author names were greyed out, and colored if selected. Should we adopt the same design? (I'm a bit mixed about it TBH. On the one hand, I like the aesthetics of the original. On the other hand, having explicit checkboxes might be a good thing for UX?)

• Expand/collapse global nav
• Adapt content pages
• Decide what to do about aside on homepage
• Stack viz controls vertically
• Prevent scroll trap on maps
• Revise mobile author selection (grid?)
• Fix medium-sized nav issue (already stashed locally)
• Make author selection collapsible

Another difference between the original design and our implementation: in the original design, dot-markers in the Trajectories view were scaled by the number of waypoints at this place. Should we replicate this behavior?

Once you zoom out all the way, you can't zoom back in. Tried in different browsers.

The original app had an absolutely-positioned legend with an explanation of the visualization on the intersections map. Would likely be worth porting over in some form.

dev notes:

• github workflow for python convertJSON on push to main

Original search: https://sameboats.org/search/

Search criteria:

• Author names
• City names
• Notes
• Citations

Should use the same map page component with prop for mode instead of state.

• convert my existing props by moving destructured props out of function signatures
• add @typedef entries for each
• scope all ReactElement to React.ReactElement
• document arrow functions (or ignore them with eslint-disable comment)

/**
 * @typedef {Object} Props
 * @property {string} prop1 - The first property of the object.
 * @property {number} prop2 - The second property of the object.
 *
 *  MyFunction takes an object of type Props.
 *
 * @param {Props} props - the props
 */

Similar to original, on click: https://sameboats.org/intersections/

• Sidebar should show each person's name and the inferred time span shown
• Either 3 levels of opacity, or opaque vs transparent date depending on date certainty

Quick question. I hadn't noticed until now, that the tiles are not showing on the project in its new home. Before I start digging into the project files to figure out why, or write the server admins to see if they have some blocks that would do that, do you have any idea why this would be, Ben? I thought you might know.

http://sameboats.org/#/trajectories

@rsimon I've assigned you since you mentioned you were planning on working on it, but let me know if you'd prefer I do this in Python instead.

should run on commit, PR

Doesn't need to follow original exactly.

• Top navigation instead of side
• Slight differences for visualizations:

Trajectories:

Floating for intersections:

• Trajectories
• Trajectories second layer
• Intersections
• Intersections data sidebar

This will be a map, with data overlays, showing one or more persons' known or possible presence in a place at a time. Records will be selectable individually, and filterable by time.

Working ideas for map/dataviz frameworks include:

• MapLibre
• Deck.gl
• React-Map-GL

Currently, the search state (=selected authors and time filter) resets when the user changes between Trajectories and Intersections pages. I played with this a bit, and the reason is the following:

• For each page change, the mapLoader function is triggered, causing the data files to get re-loaded
• The ITSBGraphProvider creates a new graph whenever the data changes here.
• Because the graph has changed, the ITSBSearchHandler will reset the search here.

A quick workaround would be to force the ITSBGraphProvider to only ever create the graph once, even if the data changes. E.g.:

useEffect(() => {
  const { authors, places, itineraries } = props;

  // was: if (authors && places && itineraries) { ...
  if (!graph) {
    const g = new ITSBGraph();
    g.init(authors, places, itineraries);
    setGraph(g);
  }
}, [props.authors, props.places, props.itineraries]);

I tested this, and it will retain state across page nav.

Caution: this would mean that the app won't function properly if a different dataset is loaded after the graph was created.This is never the case in our application, however. A cleaner solution, ultimately, would be to avoid re-triggering the mapLoader somehow.

Several cities appear in the wrong locations.

1. Go to intersections: https://performant-software.github.io/itsb/intersections
2. Click some authors.
3. Resize the window.
4. Points appear in wrong locations.

UPDATE

Maybe not a huge issue; appears that if/as you re-resize the window, the points eventually end up closer to where they should be, or exactly where they should be. Still, would be good to know what's up here.

@elotroalex I'm creating this issue you let you know about a few warnings and errors that show up during the CSV conversion. The CSVs are in the /script/in directory.

Maryse Condé (14): This entry has no key dates listed and has been skipped.
Maryse Condé (22): This entry has no key dates listed and has been skipped.
Maryse Condé (31): This entry has no key dates listed and has been skipped.
Maryse Condé (32): This entry has no key dates listed and has been skipped.
Maryse Condé (33): This entry has no key dates listed and has been skipped.
Maryse Condé (34): This entry has no key dates listed and has been skipped.
Maryse Condé (35): This entry has no key dates listed and has been skipped.
Maryse Condé (36): This entry has no key dates listed and has been skipped.
Maryse Condé (37): This entry has no key dates listed and has been skipped.
Maryse Condé (38): This entry has no key dates listed and has been skipped.
Maryse Condé (39): This entry has no key dates listed and has been skipped.
Maryse Condé (40): This entry has no key dates listed and has been skipped.
Katherine Dunham (147): This entry has no key dates listed and has been skipped.
Katherine Dunham (148): This entry has no key dates listed and has been skipped.
Katherine Dunham (149): This entry has no key dates listed and has been skipped.
Katherine Dunham (150): This entry has no key dates listed and has been skipped.
Katherine Dunham (151): This entry has no key dates listed and has been skipped.
Katherine Dunham (152): This entry has no key dates listed and has been skipped.
Katherine Dunham (153): This entry has no key dates listed and has been skipped.
Katherine Dunham (154): This entry has no key dates listed and has been skipped.
Katherine Dunham (155): This entry has no key dates listed and has been skipped.
Katherine Dunham (156): This entry has no key dates listed and has been skipped.
Katherine Dunham (157): This entry has no key dates listed and has been skipped.
Katherine Dunham (158): This entry has no key dates listed and has been skipped.
Katherine Dunham (159): This entry has no key dates listed and has been skipped.
Katherine Dunham (160): This entry has no key dates listed and has been skipped.
Katherine Dunham (161): This entry has no key dates listed and has been skipped.
Katherine Dunham (162): This entry has no key dates listed and has been skipped.
Katherine Dunham (163): This entry has no key dates listed and has been skipped.
Katherine Dunham (164): This entry has no key dates listed and has been skipped.
Katherine Dunham (165): This entry has no key dates listed and has been skipped.
Katherine Dunham (166): This entry has no key dates listed and has been skipped.
Katherine Dunham (167): This entry has no key dates listed and has been skipped.
Katherine Dunham (168): This entry has no key dates listed and has been skipped.
Katherine Dunham (170): This entry has no key dates listed and has been skipped.
Katherine Dunham (171): This entry has no key dates listed and has been skipped.
Katherine Dunham (172): This entry has no key dates listed and has been skipped.
Katherine Dunham (174): This entry has no key dates listed and has been skipped.
Katherine Dunham (176): This entry has no key dates listed and has been skipped.
Katherine Dunham (177): This entry has no key dates listed and has been skipped.
Katherine Dunham (178): This entry has no key dates listed and has been skipped.
Katherine Dunham (179): This entry has no key dates listed and has been skipped.
Katherine Dunham (180): This entry has no key dates listed and has been skipped.
Katherine Dunham (181): This entry has no key dates listed and has been skipped.
Wifredo Lam (10): This entry has no key dates listed and has been skipped.
Wifredo Lam (11): This entry has no key dates listed and has been skipped.
Wifredo Lam (14): This entry has no key dates listed and has been skipped.
Wifredo Lam (16): This entry has no key dates listed and has been skipped.
Wifredo Lam (23): This entry has no key dates listed and has been skipped.
Wifredo Lam (30): This entry has no key dates listed and has been skipped.
Wifredo Lam (39): This entry has no key dates listed and has been skipped.
Wifredo Lam (43): This entry has no key dates listed and has been skipped.
Wifredo Lam (44): This entry has no key dates listed and has been skipped.
Wifredo Lam (53): This entry has no key dates listed and has been skipped.
Wifredo Lam (54): This entry has no key dates listed and has been skipped.
Wifredo Lam (55): This entry has no key dates listed and has been skipped.
Wifredo Lam (59): This entry has no key dates listed and has been skipped.
Wifredo Lam (61): This entry has no key dates listed and has been skipped.
Wifredo Lam (62): This entry has no key dates listed and has been skipped.
Wifredo Lam (63): This entry has no key dates listed and has been skipped.
Wifredo Lam (76): This entry has no key dates listed and has been skipped.
Wifredo Lam (77): This entry has no key dates listed and has been skipped.
Wifredo Lam (78): This entry has no key dates listed and has been skipped.
Wifredo Lam (79): This entry has no key dates listed and has been skipped.
Wifredo Lam (81): This entry has no key dates listed and has been skipped.
Wifredo Lam (83): This entry has no key dates listed and has been skipped.
Wifredo Lam (97): This entry has no key dates listed and has been skipped.
René Depestre (73): This entry has no key dates listed and has been skipped.
George Lamming (55): This entry has no key dates listed and has been skipped.

There are also two warnings in the app frontend that read "skipping unconnected nodes." These are two waypoints with the following two missing place IDs:

René Depestre (11): Place ID  not found in master Place Name list.
René Depestre (81): Place ID ougadogou_burkina_faso not found in master Place Name list.

Finally, this last set does not affect app function, but worth noting anyway:

country code not found for Horta, Azores.
country code not found for Boma, Kongo Central, Democratic Republic of Congo.
country code not found for Matadi, Kongo Central, Democratic Republic of Congo.
country code not found for Mbeni, Democratic Republic of the Congo.
country code not found for Lubumbashi, Democratic Republic of the Congo.
country code not found for Kinshasa, Democratic Republic of the Congo.
country code not found for Bournemouth, England.
country code not found for Cambridge, England.
country code not found for Falmer, Sussex, England.
country code not found for London, England.
country code not found for English Provinces, England.
country code not found for England.
country code not found for Liverpool, England.
country code not found for Manchester, England.
country code not found for Hull, England.
country code not found for Margate, England.
country code not found for Ramsgate, England.
country code not found for Brighton, England.
country code not found for Lake District, England.
country code not found for Southampton, England.
country code not found for Newcastle upon Thyne, England.
country code not found for Knaresborough, England.
country code not found for Europe.
country code not found for Abidjan, Ivory Coast.
country code not found for Grand-Bassam, Ivory Coast.
country code not found for Malay Federation (Malaysia).
country code not found for Brazzaville, Republic of Congo.
country code not found for Saint Maarten.
country code not found for Saint-Thomas.
country code not found for Scotland.
country code not found for Glasgow, Scotland.
country code not found for St Maarten.
country code not found for Basseterre, St. Kitts and Nevis.
country code not found for Ho, Trans-Volta Togoland.
country code not found for Trinidad & Tobago.
country code not found for Yugoslavia.
country code not found for Bingerville, Ivory Coast.
country code not found for Portsmouth, England.

EDIT: Ouagadougou has been fixed.

Error with the python script producing the JSON.

This input type is not supported in current versions of Firefox and Safari, so needs validation to ensure month/year combinations are correct.

• Validate end date is after start date

I don't think it would cause much overhead. It would essentially be a single group-by operation, which should be quite fast. (Itineraries are lists of waypoints. You'd just need to loop through the waypoints once to get a list of distinct place IDs. Probably hardly noticable on today's hardware, even with many waypoints.) Once you have the place ID, you can map the IDs to GeoJSON features by retrieving them from the in-memory graph. Should also be very fast.

Originally posted by @rsimon in #4 (comment)

This is an issue since the server isn't configured to redirect all of these requests to index.html. Use HashRouter instead of BrowserRouter: https://reactrouter.com/en/main/router-components/hash-router

The website is a rebuild of https://sameboats.org/. It should be built following core principles of minicomp, consistent with using mature, widely deployed tech. It should have sections for:

• Navigation and layout
• Trajectories/Intersections (integrate with react-router)
• Home, Credits, Instructions
• Search
• Design for mobile

Candidates for a framework could include, but are not limited to:

• Jekyll
• Next.js
• Create React App
• Remix.run (which tells compelling accessibility and progressive enhancement stories)
• Hugo (go)
• Gatsby (js)
• Etc.