Tinderbox Forum

Using tinderbox for investigative research

Hello! I am looking for a tool to aid my research for an investigative story. I have people, companies, actions, and dates of those actions. So, I might have [Person A] [invested in] [Company B] [on April 1, 2018].

I am looking at Tinderbox for this (among other tools like Brain, Devonthink, and so on). I can obviously see how to use notes for Person A and Company B.

What I can’t figure out is how to display the action, and the date it took place. I can do a link, but the only way to have the link show [invested in] is to assign it as a link type, and that way seems to me to be a long-term disaster. (Link types for every type of action.)

And, I can’t figure out how to add a date to it, either.

Do I need to make actions into notes, with a date attribute? Then link through the action? That seems pretty busy, especially since one person might have tens or hundreds of actions I’m trying to track.

I’ve thought about doing this with some sort of wiki tool, but I really like the ability to explore relationships visually.

If you have suggestions, I’m certainly open. And if you have other software I should look at, let me know. Thanks!

This thread might have some useful tips:

and this blog article (quoted in that thread):

I suggest defining attributes for the people, companies, actions and dates. For the actions, you might want to use a set attribute with and define the allowable actions that can appear in that set up front.

Finally, as discussed in the thread, above, consider using the Attribute Browser view for your analysis work.

Welcome Bruce,

In addition to Tinderbox, you might want to have a look at Notion. It’s database centric, but I have found it to be very useful when conducting investigative research. I like the way it scales, its use of Markdown, and its ability to embed PDFs. Access of data can be gained via the web or the desktop. By creating different views into the data, the visual possibilities are almost endless. It also supports links. There is no mapping capability like Tinderbox’s, but there are some visual tools/views that I have found to be very powerful. I will continue to use Tinderbox for conceptual modeling but Notion is just another tool that I recently added to my workflow. Finally, it can be set up to be private or collaborative.

I’d suggest that you step back and think, generally, about how you imagine using your notes later in your investigation.

Let’s imagine, just for instance, that you’re dealing with a big product liability case. (This was the subject of a Tinderbox Weekend presentation a few years back.). So, you’ve got lots of depositions by lots of people involved in developing and marketing the product, and by various people who are claiming various injuries. You’ve got cases proceeding in different courts — and each of those courts has its own judge who issues their own rulings — rulings that may differ from one case to another. And nearly everyone who is being deposed, or filing suit, has their own counsel.

So: first thing: you’ve got an awful lot of people to keep track of. When the witness in a deposition refers to a memo from (say) Mary McGillicudy, you’re going to want your notes to be able to find Mary McGillicudy quickly. (This was what the attorney giving the talk said was the big win: he became the expert in his firm about who was who, and who was representing them.)

Now, there are all sorts of other things that you might want to know next year from the notes you’re talking now. Maybe you’ll discover some conspiracy. Maybe some entirely new technical issue is going to arise. Maybe it will turn out to be terribly important to know exactly when everybody knew that the product would be made in China and not in Manilla. You can’t plan in advance for that: Tinderbox works hard to make it easy to add formalization to your notes later, as you discover what matters.

So, maybe you start with a sort of annal — a list of people and things they did:

. Mary McGillicudy
10/15/17: memo to Dr. F
10/18/17: flight to Hong Kong
10/21/17: visit to Bulgarian consulate

Or, maybe each of these is its own note, with a $StartDate that contains the date of the action and an attribute $Who that contains the person (or maybe list of people) involved.

1 Like

I also would urge you to reconsider how you structure things. Based on what you describe, it reminds me of a project I did to simulate the effects on market indices (e.g. FTSE100) of various corporate actions.

I suggest you make the center of your thinking the events (or actions). Each event could have attributes including the date (or interval), the actors, the type of event, and perhaps other metadata relating to specific event types (e.g. an acquisition event type might have the purchase price).

At that point, you may be able to derive the list of companies using an Agent that collates all the actors from your events. Or you can filter what you’ve got by, e.g., using the event types with the Attribute Browser.

However, another approach is to create a container for actors, using the name of the actor as the name of the note (since names are probably unique in your namespace). With that approach, you can then programmatically retrieve the actors from an event by using their names (which is in the event’s attributes) as part of a path specification into the container. By using notes you can store other arbitrary information about the actor in the note text or other attributes.

I find links very appealing in part because I like mind maps. However, my experiences with Tinderbox have suggested to me that the places where a link is the best solution are very few. I find it better to “infer” links from the structure and content of the data.

1 Like

The other suggestions you’ve received here are valuable. Also, there is no substitute for prowling around in the software and finding the approach that matches your styles of working and thinking.

Because I find it so incredibly valuable, for my own working- and thinking- styles, and because I think some people have not been fully clued it about its power, I direct your attention to the Attribute Browser.

If you start with this thread, you’ll see a lot of examples and links. (I am a journalist and do lots of research projects, and find both TB in general and the attribute browser in particular to be indispensable.)

I think it is a mistake to think that you need to define categories before you start work. That might be necessary with other programs, but with Tinderbox I feel it is better to start work and allow the categories to become apparent while you are working. If you start by defining the categories you have effectively done some of the analysis before you have seen the data. That will shape your approach and may lead to all sorts of assumptions that you might not make if you start with no categories in mind, and allow the data to show you what is important. You might find categories that cut right across people/places/events etc.
I’ve used DEVONthink a lot, and it is more for storage and finding than for analysis, in my view. Very, very useful program, but for a different thing.


I agree with that, but Tinderbox is limited, IMO, when the data to analyze are documents. Importing documents into Tinderbox has limitations, and when a “sizeable” archive is involved (more than a few dozen) things start breaking down. Obviously, this is dependent on one’s specific situation. Sometimes it is necessary to employ a two-step process and use DEVONthink to extract relevant bits from documents, then import (via the “watch” feature) those bits into Tinderbox for the complex analysis.

Quite right, in my view. That is basically what I do.

Thanks to all for your replies! I have read them all, and am going to spend some time exploring your suggestions.

One of the key comments above, I think, is the use of DevonThink vs TinderBox. I’ve looked at DT more than once, but its file-based approach doesn’t seem suited to analysis. On the other hand, TB seems well-suited to linking things, and finding un-obvious relationships and other possible avenues of exploration.

Perhaps I can use TB as a “front-end” to the longer notes themselves – as well as a web of persons, events, and organizations in my research space.

Exactly: DEVONthink and Tinderbox are different beasts. Lots of people need both!

They work well together.

DT is godsend when it comes to storing stuff. Sure, you can do that in Finder, but once you’ve tried DT it’d be hard to go back, because it makes sorting, classifying and digging through your stuff a lot easier. And as @eastgate said, it works well with Tinderbox.

For me (and depending on what your ultimate needs are) the holy trinity are DT, Tb and Scrivener.

1 Like

DT 3.0, released today, expanded its support of Markdown to the extent that you can now use Wiki links. Good stuff.

Just be aware that so-called “wiki links” inside DEVONthink are not hyperlinks, and are managed only within DEVONthink. That is, they would not work on export.