The Distant Librarian

Have you trained an AI?

Thanks to Mita Williams for pointing to this Washington Post article that makes it trivial to search and see whether any sites you're affiliated with have been used to train "Google’s C4 data set, a massive snapshot of the contents of 15 million websites that have been used to instruct some high-profile English-language AIs, called large language models, including Google’s T5 and Facebook’s LLaMA." The entire article is interesting, but it's more fun to plug in the URLs of any websites you want to check against. This blog, and MPOW, are both represented:

Is anyone aware of a similarly-simple way of searching the entire Common Crawl dataset, or is that just too massive to do?

If you're curious to know whether any of your images or artwork have been used to train one of the image AI's, you can do that, too!

 

 

Clipchamp - a VERY cool screencasting option!

The other day, Lifehacker posted that Windows Has a New Tool for Simultaneously Recording Your Screen and Webcam. That tool is Clipchamp, and I am impressed! Oh, and it's free.

MPOW has blocked the Microsoft Store, so I wasn't able to install the desktop version, but I was able to run the web version. I have to imagine the desktop version would provide a more seamless experience, with better performance, but even running it in a web browser was perfectly fine. Maybe a little jerky when doing the editing. The web version requires either Chrome or Edge, but running it in Chrome, I was able to record what was happening in my Firefox browser. I'm not going to bother posting the test video I made, but it did everything I expected. When editing, I was able to clip out sections, I was able to move and resize my video camera feed, and I added some stock music and graphics. When saving, I had the option to export at different resolutions, and then have the option to upload directly to YouTube or 7 other destinations. Exports are .mp4 files.

There is a limit of 30 minutes per recording, and I don't see that the Premium version changes that. For 99% of what you're likely to want to record as a screencast, I think this tool would do the trick. Definitely worth checking out!

AI, Images, and Academic Titles

Yesterday, shortly after helping a student start a literature search, I found myself poking around at DALL-E, one of the crop of machine-learning tools that can generate images based on a text prompt. I had an article title saved on my clipboard, and wondered what would be generated from the title. I was not impressed.

The prompt: Robust Speech Recognition via Large-Scale Weak Supervision

The 4 results:

Here's a second example.

The prompt: The curvilinear effect of professional faultlines on team innovation: The pivotal role of professional identity threat.

The 4 results:

At least that one's got some curves...

Pretty underwhelming, but then I wondered what the others would make of it, so I headed over to Midjourney, and after figuring out how to interact with it in Discord, generated the following MUCH more interesting images from the same prompts. Can you guess which is which this time, w/o seeing the prompt? :-)

 

And finally, what about Stable Diffusion?

Could all turn in to a fun party game. I certainly know which ones are most visually-appealing to me... Thoughts?

Why you should pay attention to your Creative Commons licenses

I have never given a second thought to ensuring I've chosen the correct/most recent version of a creative commons license. That's changed, because I just finished reading Cory Doctorow's post, A Bug in Early Creative Commons Licenses Has Enabled a New Breed of Superpredator. In it, he outlines the emergence of the Copyleft troll, someone who posts images with an outdated CC license, and then sues people for using those images as they were intended to be used, but who may (or may not!) have neglected to completely comply by the letter of the license.

I don't know whether any of this would even apply to the textual works I would more likely post with a CC license, and maybe that doesn't matter because I'm not going to be going after anyone for using my works that may have an old license, but here at MPOW we're really starting to move to Open Educational Resources, and I could easily see the authors of such resources including images that are CC licensed earlier than 4.0 (or whatever the current version is by the time you read this). Going back to Cory's post, it seems that license version 2.0 is the really problematic one, but perhaps 1.0 and 2.5 are bad for the same or similar reasons? Please let me know if you know!

Does anyone know of a tool that allows me to search a specific site for content that has any creative commons license? So I'm looking for something like https://wordpress.org/openverse/ that includes something like Google's "site:" option. Bonus points if it could allow a search to be limited to content that applies a specific CC license! @researchbuzz? @cogdog?

Resources:

• Creative Commons FAQs
• Creative Commons version 4.0
• Creative Commons history of license versions

I love these quotes from Cory:

If you put a CC license on your work, its explicit message is, “I want you to re-use this.” Not “I am a pedantic asshole with a fetish for well-formed attribution strings.” The point of CC is not to teach the world to write attribution strings: it is to facilitate sharing and re-use. If you are a good-faith user of CC licenses, then your response to an incorrect attribution string should be a request to correct it, not a threat to sue for $150,000 in statutory damages.

If you send threats instead of requests for correction, you are a terrible person and you should feel really bad about yourself.

Stay safe, all, and do let me know if you have answers to any of my questions above!

 

Shasta vs. Descript

It's a premature title, as I don't actually have access to Project Shasta yet, but this announcement makes it sound very similar to Descript: Adobe’s Project Shasta is an AI-powered, web-based audio editor. I have requested access, though, and will report back when I know more! :-)

Linked data types in Excel

Apparently parts of this have been around for a couple of years, but it just came on my radar last week with this blog post.

Did you know that Excel can do lookups from providers like Wolfram, Bing, and Power BI on a bunch of different data types? Only available with a Live365 (and/or institutional?) account though, but that covers MPOW. Here's the overview page at Microsoft.

A partial list (you provide the first part, Excel will return data on the second):

• Food: calories, carbohydrates, fats, vitamins, nutrients, etc.
• Exercise: exercise types and calories burned.
• Locations: zip codes, economic data, airports, schools, forests, etc.
• Universities: graduation rates, tuition, student body, etc.
• Chemistry: elements, compounds, and minerals.
• Space: planets, moons, satellites, supernovas, space missions, etc.
• Movies: actors, characters, directors, release dates, poster art, etc.
• And More: music, plants, animal breeds, famous people, languages, etc.

I was going to do a little screencast, but there are good ones already out there.

You can even define your own data types.

I'm not entirely sure what I'll do with this, but at least now that I know about it, I'll keep my thinking cap on!

Historical COVID-19 R-Values for Alberta

Filed under stuff no citizen should have to do, today I went through pages archived by the Wayback Machine in order to provide a patron with historical R-Value data for COVID-19 for the Province of Alberta. This information is updated weekly (since mid-December, 2020) at https://www.alberta.ca/covid-19-alberta-data.aspx, but for some reason the Government has chosen not to make previously-reported numbers available. Anywhere.

Here are the Historical COVID-19 R-Values for Alberta in a Google Spreadsheet. I'll continue to manually update.

 

That cool picture from the cover of The Data Journalism Handbook

The Data Journalism Handbook has been updated from its original 2012 edition, and is available to purchase or for free download from the publisher. When I first tweeted about this a month ago, @Gersart wondered who was responsible for the artwork on the cover. Frustratingly, there was no information about it in the publication, and my email to the publisher went unanswered.

So I clipped the image and threw it into TinEye and was able to identify it as a photograph from a 2012 exhibition by Sarah Sze entitled Fixed Points Finding a Home.

See also, and this YouTube video.

Bounegru, L. and J. Gray (eds.), The Data Journalism Handbook: Towards a Critical Data Practice.
Amsterdam: Amsterdam University Press, 2021

How to stop a Zoombombing

Thank goodness I haven't personally experienced this while teaching / presenting via Zoom, but I have been a participant in a workshop that was Zoom bombed, and that was traumatic enough. I can't even imagine how nasty that would be for the presenter.

Lifehacker reports that Zoom is rolling out some new features aimed to make it easier to stop a bombing-in-progress, but also appears to be monitoring for, and will warn you via email "if your meeting invite appears in any public social media posts or other websites where these things are usually posted."

I haven't received anything from my POW (we have a licence for Zoom), or from Zoom directly, nor did I see any abuse options as described by Lifehacker until I manually checked for updates in my Zoom client and updated to version 5.4.3 (Mac).

Go update your clients, folks!!!

Preparing a dataset to use in Hydrator

Update - April 18, 2023: From the developer: Twitter's changes to their API which greatly reduce the amount of read-only access means that the Hydrator is no longer a useful application. The application keys, which functioned for the last 7 years, have been rescinded by Twitter.

Every few months I get an email from someone looking for some help with the Hydrator tool I wrote about back in 2017. Hydrator is a simple tool to rehydrate twitter ids into full tweets. From the Hydrator GitHub page, "Twitter's Terms of Service do not allow the full JSON for datasets of tweets to be distributed to third parties. However they do allow datasets of tweet IDs to be shared. Hydrator helps you turn these tweet IDs back into JSON and also CSV from the comfort of your desktop."

The advantage for me, and apparently many others, is that it's an actual program instead of a command-line tool, so it's pretty damn easy to use for those of us who only occasionally need to rehydrate a list of tweeter ids.

The biggest issue people have is that Hydrator only wants a file with a single list of twitter ids, and nothing else. Not even a header on that list. For a reasonably-sized list it's usually no problem to use Excel or similar to clean up the offending dataset, but the other day someone asked if I could help with prepping the file at A Twitter Dataset of 150+ million tweets related to COVID-19 for open research. For some reason the researchers added two columns to the dataset, so in addition to the column of twitter ids, there are also columns for date and time.

Note this is all happening on my MacBook. YMMV on other platforms.

I did some Googling and decided that csvkit should do the trick and it did, with a couple of caveats. First, it's an incredibly simple command-line tool, with a stellar step-by-step tutorial. Trust me, you can use it! So I installed it and ran the first command to turn the COVID-19 dataset from a TSV (tab-separated values) file into a CSV (comma-separated values) file:

in2csv full_dataset-clean.tsv > full_dataset-clean.csv

And it didn't work :-( csvkit is supposed to be able to recognize and parse tsv, but for whatever reason it didn't in this case. No biggie, back to Google, and after a little trial and error, I settled upon this single command line to convert the file from TSV to CSV:

cat full_dataset-clean.tsv | tr "\\t" "," > full_dataset-clean.csv 

And just below that is a suggestion that I might not've even needed csvkit for the next step, but I didn't see that until just now :-)

So that command did indeed translate my TSV file into a CSV file, and I went back to the csvkit tutorial to get a file with only the first column, using the command:

csvcut -c 1 full_dataset-clean.csv > full_dataset-clean-tweet_ids.csv

One last problem is that there's still a header on that column (tweet_id), which I verified with the command:

csvcut -n full_dataset-clean-tweet_ids.csv

Google suggests the SED (stream editor) command will do the trick:

sed -i -e '1d' full_dataset-clean-tweet_ids.csv

And I can verify the first line is now gone with:

head full_dataset-clean-tweet_ids.csv

Probably way more convoluted than it needed to be, but it got the job done, and now I can point Hydrator at my full_dataset-clean-tweet_ids.csv file and Bob's your uncle!

Note to self, xsv is another CSV tool that might be worth exploring.

Photoshop 2020 blank screen instead of images - fixed

Just had the pleasure of upgrading to the latest version of Creative Cloud at MPOW, and while it appeared to go smoothly I immediately had a problem with my workflow. Importing an image (jpg) into Photoshop would only show a blank white screen. The image WAS correctly rendered on the home screen, but just a white workspace when attempting to actually work with the image.

I did some googling and found a similar problem reported with a fix for the Mac version, and was able to figure out where the same settings were in Windows, and it fixed my problem.

So for my future reference, and maybe yours, go to Edit, then Preferences, then choose Performance. Uncheck the box for Use Graphics Processor. This was on a Dell OptiPlex 7050 running Windows 10. Good luck!

How to Find the RSS Feed URL by Guessing

The Zapier newsletter I just got had a link to this article from last year: How to Find the RSS Feed URL for Almost Any Site. It's got some good tips for sussing out RSS feeds, but I have a couple more.

I ran across two sites recently that should've had RSS feeds, but apparently didn't, at least according to Inoreader, my feed reader of choice. Inoreader usually does a good job of automatically detecting an RSS feed if it exists, but it didn't find one for www.canadianenergycentre.ca^*, which is a WordPress site. Turns out if you simply add /feed to the end of the URL you'll be good to go. That may be true for all WordPress sites - worth a try at least. Not sure why Inoreader didn't find that, and I'm still waiting to hear back from their tech support.

The other site that's crying for a feed is The Far Side! I emailed to ask for one but never heard back, so I built one of my own using Feed43. I'm not posting it here for fear of it being removed, but you can build one yourself, or hit me up privately for the URL :-)

Long Live RSS!

^*This site doesn't get any Google juice from an actual link from me

3 Simple, Free Apps for Accessing Your Home Computer From Anywhere (Gizmodo)

AnyDesk is totally new to me, so I thought I'd make note of Gizmodo's post on 3 Simple, Free Apps for Accessing Your Home Computer From Anywhere. I use Chrome Remote Desktop at least weekly, and have used TeamViewer to good effect in the past, so can vouch for each of those. The commenters overwhelmingly choose TeamViewer. I'll certainly explore AnyDesk and likely add it to my arsenal.

Good tools to know about for sure!

A quick showdown of three automatic transcription tools

One of my hats these days has me supporting students and faculty in the use of NVivo, a piece of software used for qualitative data analysis. Often people are wanting to analyze the text of recorded interviews, and of course that usually requires someone to transcribe the audio into text. I'm always on the lookout for free tools to automate this process.

I didn't realize it was that long ago, but last December I started playing with an automatic transcription tool called AutoEdit2, and found it pretty decent. Yesterday and today, ResearchBuzz led me to two new options, so I thought I'd do a quick comparison.

First up was a link to this announcement on TechCrunch about Deepgram. Then today was a pointer to Hongkiat's How to Transcribe YouTube Videos Automatically, which actually has pointers to several different tools or methods. I was most interested in the new-to-me option of having Google Docs automatically type what it heard into a document. Do note this method requires the use of the Chrome browser - I missed that and couldn't find the option in Firefox when I first tried...

For testing purposes I grabbed a short audio clip of George W. Bush's Oath of Office from The George W. Bush Public Domain Audio Archive.

Here's what Deepgram made of it:

please your right and never after me i george walk bush too solemnly swear i told what university so this work but i will say for extra office of person of the united states that i was lately excuse a of the united states and roll to the best of my and well the best of my ability sure protect learn the constitution of the united states sarah protect and the fan constitution of the united states saw not like so help

me god in congratulations

Here's what Google Docs heard:

The only answer please raise your right hand and repeat after me I George Walker Bush do solemnly swear a George Walker Bush to sell them I swear that I will Faithfully execute the office of President of the United States that I will Faithfully execute the office of President of the United States and will to the best of my ability and will to the best of my ability preserve protect and defend the Constitution of the United States Constitution of the United States

And here's what AutoEdit2 came up with:

the only please raise your right hand and repeat after me I George Walker bush do solemnly swear I George Walker bush do solemnly swear that I will faithfully execute the office of president of the United States that I will faithfully execute the office of president of the United States and will to the best of my ability and will to the best in my ability preserve protect and defend the constitution of the United States preserve protect and defend the constitution of the United States so help me god so help me god congratulate

So it seems to me on this very short audio sample that AutoEdit2 did very well, followed by Google Docs, with Deepgram bringing up the rear. Deepgram is the easiest to use, requiring the user to simply upload an audio file or point to either a YouTube video or online audio file. Once installed (MacOS only), AutoEdit2 is also very simple, though does require the file to be local, not online. And Google Docs is a bit of a pain to set up initially, but after a little trial and error is fairly straightforward and can record either a local file or something streaming online.

Curious to hear if you've played with any of these, or other tools, with longer audio clips, and if you find the same results?

Search Google Maps via shortcuts

Lifehacker has a post describing how to Map Anything From the Chrome Address Bar. Sounds great, except I recently switched back to Firefox after many years away, so found that there are a pair of extensions for FF that allow you to do pretty much the same thing. Map with Google Maps takes care of the highlight searching, and GoogleMaps adds Google Maps to your list of search engines, allowing you to map addresses w/o starting at the site itself. And if you prefer Open Street Maps...