punishments for not meeting goals?

& stop partnering with firms that turn around and monetize/gatekeep that data?

On which note the gov't needs to talk to the webarchives.ca etc team, researchers in this area

one stop portals - hmmnm. better to have a well documented API, with 'recipes', tutorials and guidance on how to query it.

This should be an extremely high priority indeed.

The government needs to look at the lessons learned in the US, especially via https://18f.gsa.gov/

http://www.fastcompany.com/3046756/obama-and-his-geeks

Will this website be free of tracking or analytics? Right now, an ATIP request requires a simple email. Why not ditto for personal info?

I have an outstanding ATIP request that is almost exactly 2 years old now. Will the Government hire more people to process the request? Nothing I see here does anything to speed the process up. How does the government propose to deal with footdragging within a department reluctant to fulfil a request?

Part of 'access' must involve education, that is to say, to address the 'so what?' question. What shall people do with this data? A csv dump is fine: how do you educate citizens about what to do with this data? Data hides as much as it reveals.

Part of 'access' has to address the end-point. The distribution of access to the web is patchwork in this country, especially in rural areas. Will this mean things like support for local libraries? Despite years of acknowledgement of the importance of access to the web, we are still locked with telecoms that make access prohibitively expensive.

Without provenance, there is absolutely no reason whatsoever to believe anything King says about the parchment.

The lack of interest in provenance is what enables the trade in illicit antiquities to continue unabated.

6600 files were downloaded; 333 files appear to be these missing editions with the placeholder text. I have not yet manually verified all of this... which is partly the point, right?

6267 print editions, from 1893 - 2010

that is, from .72 to .9. They all have the same placeholder text, but the quality of the ocr makes some consistent errors, which is interesting.

welp, here's some of my stuff http://shawngraham.github.io/presentations/slippage.html#/

contrast with: https://hyp.is/AVSqo9ZlFtPwhMO7Dnqf/www.ctheory.net/articles.aspx?id=519

graph all this...

I need to carefully take a look at the txt, and see what other regex patterns need to be added

Observation: Paul Reilly

Observation: additive 4 right. b/c that would cause us to rethink the ways we break apart / build up archae knowledge

Observation: additive3 printing out site in material that replicates literal granularity of context description; haptics, sonic, other qualities Resonance: Stu Eve's embodied GIS, phenomenology Crossref: JAMT 19 (4) pp. 582-600

Observation: keytheme2 The idea of 'generative' works on many levels here.

read this

Grand disciplinary challenge

Instability of Reality/Virtuality

Additive Manufacturing

.

check this out.

-go back and remind yourself what the 3 stages in evolution of additive tech are - shape, to control of multiple materials, to control of programmability, what?

in line 43, that's where you put the link to the page you actually have annotated. fyi. forgot to write this down.

.

tadag

ontology

theme

Things start getting very inception like when I annotate the page displaying the annotations.

This is where DH comes into play, too .

this is where digital humanities intersects with machine learning, perhaps.

@sorcha on twitter https://twitter.com/_sorcha/status/654787074273316864 writes: "I'm not convinced by your dismissal of 'window dressing' when the mimetic experience of historical games is such a big draw"

"...soundbashing gets us closer to the original messiness of data, and in its weirdness, undermines the authority of data visualization."

I've tried others; search Electric Archaeology for 'open notebooks'

Link is broken.

Sometimes, it doesn't. It will be running, but you will need to point your browser to this address:

127.0.0.1:3333

The 127.0.0.1 is your home machine (also known as localhost, while the :3333 is the port that open refine is serving up at.

typo. should be cpan

is that an 'el' or a one? damn typography!

This is really quite useful here: http://biohackathon.org/d3sparql/ you can feed it a query and get a visualization of your results, eg

Some of my students were finding (chiefly on Windows machines) that they needed the -m flag, eg

python -m pip install pymarc to call pip successfully.

Give this a try if you're on windows https://sites.google.com/site/pydatalog/python/pip-for-windows

We've found that we get an error message to the effect that we do not have permission to write. See https://twitter.com/electricarchaeo/status/696703569337176064

this line is throwing an error on Windows 10 systems, to do with character encoding.

UnicodeEncodeError: 'charmap' codec can't encode character

From googling around, this seems to be an issue in the console window for the command prompt. This thread might be useful towards a solution: http://stackoverflow.com/questions/32694994/utf-8-on-windows-python

this needs to be indented

Something that would be nice to have is perhaps a suggested progression or path through the tutorials towards different end goals.

For the complete beginner, I'd suggest the 'getting started with bash' lesson.

Then, in other lessons, keep an eye out for assumptions about whether or not you're to work at the command line (windows), the terminal (mac/linux), whether or not you're supposed to invoke python etc.

What kind of people? Historians? Clearly. But given the kind of people you were discussing at the beginning of the article, I wonder if you can expand on the idea of perhaps citizen-scholars doing their own social media analysis on hashtags, events, of importance to their own communities. An opportunity for a kind of public-humanities/sociology? An opportunity for a more informed citizenry? Imagine for instance if some new social movement sweeps the country, and its activists take the time to analyze their own hashtag/presence on twitter, would it (could it?) change their tactics, strategy, outreach, etc....? just a thought.

Great piece fellows! Look forward to seeing this coming out.

'twarc', a tool developed by Ed Summers ... etc?

Seems to me that you could extract from this paper an excellent Programming Historian tutorial in short order.

more?

This is a fascinating aspect to your data. I wonder if you can elaborate more on this, because surely this archival silence is a story in waiting. Does twitter eat the loser?

Extremely important. This point should perhaps be made more forcefully for the benefits of the historians amongst us who are dismissive of social media and its utility as a historical primary source...

Ethics of highlighting usernames here?

great to have this. what story do these most rt'd tweets tell? the four word clouds are brilliant at capturing the transition (although on election day, given that poll results were being made public in the east before polls closed in the west, i wonder if those tweets cease to be about 'the election' (the process) as the result. just a thought.

The first four of the most retweeted tweets are from election day itself, projecting results etc (and the first one, that most photogenic of couples the Trudeaus). It might be more interesting (?) to know what the most RT'd tweets were up to the day before election day.

so that's how you do it! damn, that's awesome.

is it useful to explain what deduplicating is, or why there might be duplicate tweets in the first place? duplicates being the same as rts or something different? # of times rt'd is surely an important aspect of the data.

is it worth explaining what this line is doing? or presumably the audience is literate enough to decode this? I know I've always had a devil of a time getting twarc report to work properly, so this example is certainly handy for me.

I know that this is baked into the TWARC code, but it kinda sounds as if it's Ian and Nick who make this call. So there are ideas of good digital citizenship baked into the tool you use... which makes me wonder: if you were using something other than TWARC, would there be a significant different to your results? Or if the other tool was unethical in its treatment of the data presumably you'd be tainted by association. Hmmm. Does York and Waterloo have ethics board guidelines for social media search?

I'm still not clear on what the difference is between the search api and the streaming api. Is there a way you can highlight the difference in a concrete way?

oh noes! I wonder how this plays into the analysis subsequent...

strikes me that this phrase is something that needs unpacking. How does twitter define 'relevance'? Simple keyword matching?

$50k US knocks a lot of Canadians out of the pool, of course...

sentence fragment?

Given that something approaching 50% of Canadians live in the five major urban regions, and internet connectivity & use in Canada is historically correlated with tertiary level education (http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/comm35a-eng.htm), isn't there a chance here that 'every day people' might be a bit misleading? That the tweets map against the same demographic who read the Globe and Mail rather than the Sun, as it were?

(I know that table is out of date, and perhaps there is better statscan data out there now. Perhaps this isn't an issue anymore. But it was a factor in our study from a few years back, so I thought I'd raise it here http://quod.lib.umich.edu/d/dh/12230987.0001.001/1:9/--writing-history-in-the-digital-age?g=dculture;rgn=div1;view=fulltext;xc=1#9.3)

An impressive number, to be sure. Does that represent every tweet made during that period? I just ask because of Ed Summers' recent post highlighting changes to the Twitter API http://inkdroid.org/2016/01/28/api-studies/ and I wonder what the API was doing during the election.

What does 'exponential increase' actually mean though? In comparison to what? Isn't there research out there that gives a sense of what Twitter's demographic base is? Given that you've got the twitter stream data, and that some kinds of demographics can be inferred from the metadata (location, operating system / device as proxy for relative wealth etc)...?

Pymarc depends on someting called 'six'. If Six is not up to date, you can get an error, cannot import name python_2_unicode_compatible when you run import pymarc in the python interpreter or in your .py script.

To solve this, try running 'sudo pip install --upgrade six`. On my machine, the six version was 1.4.1; upgrading brought it to 1.10 and pymarc worked.

Windows: 'python -m pip instal --upgrade six'

This link is out of date- the get-pip.py script is now at https://bootstrap.pypa.io/get-pip.py

In any event, Python 2.7.9 or greater has pip in it already, as does Python 3

Yes! Yes! A 1000 times, yes! Or at least, to the questions in the second line, right?

I'm hoping users of Hypothesis will find this workbook helpful, and that they will annotate it accordingly.

can I annotate here?

Testing hypothesis on a hypothesis

This is me leaving an annotation