My content now lives on my site – more than ever, that is. Over the past few weeks, I’ve created a new space on this website where I can keep a collection of almost “everything” I’ve ever published online, and one where I can write journal-like “notes.” Here’s what to expect and why I did this.

Everything: All there is from me

I love putting stuff online. Since I was 16 (that’s 20 years ago), I’ve been doing so – on now-dead forums, now-dead social networks, and surprisingly still-alive blogs. Seeing the things I created online makes them better, I find. I’ve shared photos I took from my high school at 16, work with which I applied for art school at 18, and enjoyed documenting my master’s thesis (2014) and a fellowship (2016) almost daily. And over the last 16 years, I’ve learned to love and then mourn Twitter.

For these ~1,100 posts from my past, I created an archive called “Everything.”

It includes:

• 493 tweets that I posted first as @lisacrost (2009–2021), then as @lisacmuth
• 204 photos from an old Blogger blog from when I was 16–18 years old
• 170 drawings from an old Blogger blog from when I was 17–19 years old
• 25 posts documenting an internship at Bloomberg in New York City in 2013, originally on Tumblr
• 99 posts documenting designing my Master’s Thesis in 2014 (Tumblr)
• 133 posts documenting my OpenNews fellowship in 2016 (Tumblr)
• 131 posts documenting my research for the book on colors I’m writing (>2021), originally on Notion
• 14 (so far) book reviews from Goodreads
• 12 thoughts about data vis books I read for the Data Vis Book Club, from the written discussions about those books.

I curated. Many of my 5,100+ tweets are boring (“Don’t miss our data vis meetup tomorrow!”), so I only kept the ones I want to preserve for the future. You won’t find replies or pure retweets in there, either. The same goes for my drawings and the photos I took as a teenager: I hope I hit a sweet spot between showing authentically how 17-year-old Lisa saw the world and showing stuff that’s still interesting somehow.

Notes: My little corner of the internet

On “Everything,” you’ll also find the 30 posts that I shared over the past three months in another new section of my website: “Notes.” This is where you can watch me share sometimes boring, always unpolished thoughts. I upload photos I took that day or the day before, doodles, book reviews, visual experiments, and work in progress. (Here’s a full article on how I’m using RSS to post to this Jekyll blog on my phone, if you’re interested in that.)

What I’m doing there is different from what I usually do. When I write on the Datawrapper blog and on social media, I think a lot about you: what you’ll find interesting or useful or fun. I research, structure and restructure, edit and proofread — all so that you get to read something that’s as valuable as I can create. I enjoy doing so, and it makes me really happy to know that you enjoy reading it.

“Notes,” however, is my little corner of the internet. It’s a small, cozy room of my own where I go if I don’t want to think about you for a while. Here, I write for myself. I still censor myself — this is the internet, after all, not my actual journal or best friend — but I post lots of things that wouldn’t feel right as a proper article, on the Datawrapper blog, or on social media.

That said: I won’t mind guests! Do visit my little room, or at least peek through the window from time to time. In fact, I created an RSS feed for you to do just that.

ROOTS before POSSE

Why am I doing all this? Because I got inspired by the concept of POSSE: “Publish on your own, syndicate elsewhere.” For me, ROOTS is the logical first step toward that: “Return Old Online Things to your own Site” (yes, I made this up). Why? If I do decide to delete my X account or if Blogger gets quietly discontinued, then I don’t care: it’s all on my site already. I own it. It’s all Markdown files and images that I can back up anywhere I want.

You’ll see me POSSE (or PESOS – “Publish Elsewhere, Syndicate (to your) Own Site”) in the future, too: If I post a Goodreads review, it’ll also be on my Notes and Everything pages. If I post on LinkedIn, it’ll be there, too. Everything I create and find important will eventually end up on my website.

Bonus: No more cookies for you!

If you’ve been on my website before, you might notice that something else has changed: I don’t have a cookie warning anymore. I deleted my Google Analytics tracking code. And I only show you those old Disqus comments on articles where they existed, and only if you agree to it. So: I don’t know if two people read my notes (my mom and I) or hundreds. I like that.

Head over to Everything to stroll around my posts, or have a look at the new Notes. I hope you enjoy the two new sections.

My whole website runs on Jekyll, hosted on GitHub Pages. I like Jekyll because I understand it: Every single post is a Markdown file. All images live in one folder. Templates for posts use HTML, CSS, and a simple language called Liquid which I enjoy using.

Why Jekyll isn’t great for quick posting

I’ve written many articles in Markdown on long Saturday nights in front of my computer, and it has always worked well. But it’s also a bit of work:

• There’s no drag and drop to quickly add images. You have to put images in a specific folder and reference them correctly.
• Every post needs to have a front matter that specifies e.g. the thumbnail image, tags, or the categories your post is in. Adding (or copy/pasting) this front matter is tedious.
• Your posts need to have a correct file name that includes the publishing date, e.g. “2026-03-26-note.md”
• After writing posts, you need to commit and push them to GitHub — definitely more complicated than hitting “Publish” somewhere.
• All of this means that you’re dependent on your desktop computer. Writing and publishing on mobile phones is simply too tedious.

For my new Notes section, I wanted something quick. I wanted to open an app on the go, put in some thoughts or images, hit “Publish” and call it a day.

My solution for this? RSS.

RSS and the Tumblr app come to the rescue

RSS is amazing. I’ve been a RSS reader fan since Google Reader (RIP). It’s my favorite way to stay up to date on what other people post. Almost every blog comes with an RSS feed — even my old Tumblr blog. “Would it be possible,” I asked ChatGPT one evening last December, “to check this Tumblr blog RSS feed periodically, and if there’s new content, transform it into a Jekyll Markdown file and put it on my blog?” Yes, it said, and now I have the following setup:

1. To post a blog post, I open the Tumblr app and write and publish a post.
2. It gets published to an old Tumblr blog I have. It’s set to “not indexed”, so search engines shouldn’t mention it. I also told Tumblr that I don’t want my blog to appear in their search and recommendations.
3. When I publish, the whole post content (text and links to the images) immediately becomes part of an RSS feed.
4. I set up a GitHub Action with a Python script that checks this RSS feed every six hours for post IDs it doesn’t know yet.
5. If there’s an unknown (meaning, new) post ID, it transforms the text of that post to Markdown and creates a new file in my _notes folder for it. The script also downloads the images into the right image folder in my blog setup. It then writes the post ID into an extra txt file, so that it doesn’t do the whole text and image processing for this post again the next time.
6. The GitHub Action then commits all changes.

And that’s it! Every post that lands in the _posts or _notes folder in GitHub gets published automatically. Within a minute or two after the Python script doing its thing, the post is live on /notes.

Why I like this workflow

There are lots of advantages to using the Tumblr app:

• It works on phones. Initially, I did quite a lot of research into using a Markdown-based writing app and GitHub on my phone, but opening two apps for one small note feels like a perfect excuse to just not post.
• Writing new posts looks and feels smooth. Formatting, uploading images, publishing all works as nicely as you’d imagine from a blogging app with lots of money and people behind it.
• The editing experience doesn’t force me to have a title. My notes don’t have a title, so that’s neat. (The title is the date, which Jekyll gets from the file name, which the Python script writes based on the publishing date of the Tumblr post.)
• I can still add metadata to my posts, by creating rules in my Python script. For example, I add tags by adding a new line in my Tumblr post that begins with a hash, e.g. „#Data Vis #Elections“. I told my Python script to not include such a line in the content of my Markdown file, but extract the tags in that line for my front matter. I could do the same to add categories, a summary or title, a different publishing date or layout, etc.
• I don’t get sucked into doomscrolling. I’m not following other Tumblr blogs and really don’t care about the stuff I see in the feed when I open the Tumblr app.

It feels especially neat that I’m using Tumblr for the UI, but not for hosting my content (well, only for six hours max). If my Tumblr account got closed tomorrow, my content would still live on my site. I’d just move to another blogging platform and do the same trick.

The only disadvantage I can think of is that I can’t edit posts with the same workflow once they appear on my site. (It also feels a bit weird to have your posts on the web twice.)

The alternative: An Alfred workflow

For the times when I don’t want to wait for up to six hours, I also created an Alfred Workflow. It allows me to write my blog post in iA Writer and drag images in there from wherever I want (they won’t show up, but their path will). I then save the note as „note.md“ on my desktop and run the Alfred workflow „Publish blog post!“ — which moves any „note.md“ files on my desktop to my _notes folder, renames it, adds a front matter, puts the images in the right folder, and commits and pushes the post.

And that’s it! As so often, setting up that GitHub Action and Python script was mostly possible thanks to Gemini, Claude, and ChatGPT. (I’m kind of proud that at least the idea came from me.)

For our feature announcement “New: Combine series with different intervals in line charts”. Read it here.

For our feature announcement “New: Add comparison columns for election reporting and more”. Read it here.

For our feature announcement “New: Small multiple column charts in Datawrapper”. Read it here.

And yes, I got to write a few Weekly Charts, too. Most noticably:

• We’re waiting longer than ever for our favorite TV shows to return
• A bad chart got better – and then bad again

Here’s more:

For my Weekly Chart “Where would Germany be without Fukushima?”, published on April 4th, 2024 Read it here.

For our feature announcement “New: Annotations in bar, range, and dot charts”. Read it here.

For our feature announcement “New: Automatically label data points in line charts”. Read it here.

For our feature announcement “New: Add patterns to choropleth maps”. Read it here.

For our feature announcement “New: Small multiple line charts in Datawrapper”. Read it here.

I posted this article elsewhere (most likely on the Datawrapper blog), which is why you don't see the full content here.

Which tools and approaches you can use to find, come up with, or extend a color palette.

I posted this article elsewhere (most likely on the Datawrapper blog), which is why you don't see the full content here.

Since the early 2000s, wait times between seasons have gone up.

I posted this article elsewhere (most likely on the Datawrapper blog), which is why you don't see the full content here.

...so that all readers feel respected.

I posted this article elsewhere (most likely on the Datawrapper blog), which is why you don't see the full content here.

"But is that last chart overkill? Probably."

I organized the first free, virtual Datawrapper conference from March 13–15, 2024, with 33 speakers giving 30 talks. You can learn more about it here: all talk recordings on YouTube / blog posts about all speakers / agenda as PDF.

The keynote speaker was Amanda Cox, and we could win speakers from the Los Angeles Times, El País, Axios, YouGov, Reuters, AP, The Times, Zeit Online, Spiegel, and more organizations. The conference ended with an interview of Alberto Cairo about his latest book.

Why a conference?

When starting to work at Datawrapper, I’ve quickly become amazed by what users achieve with it. We’ve showed some of their work on our blog – but to truly make these users benefit from each other, I felt like an conference was needed.

We called it Unwrapped, because that’s what it was: The conference opened up and unwrapped how others are making the most out of Datawrapper. It was a space for advanced Datawrapper users to learn from each other — and for beginners to get to the next level.

Datawrapper has a global user base, and we wanted to give everyone the chance to speak and attend, so the conference happened virtual.

Supported by the trust and advice of the Datawrapper CEOs, I organized the following:

Communication with the speakers

Unwrapped featured 33 speakers, from the United States, Germany, Ireland, Finland, Ghana, U.K., and Spain. They signed up for an experiment (we’ve never done that before!) and I was so grateful for them to do so.

• Call for speakers in December 2023. Setting up a form for the call, deciding for the speakers, and communicating with everyone who applied to speak.
• Reaching out to other potential speakers
• Communication with every speaker before and during the event: Deciding on talk topics with them, keeping them informed, organizing that they could try their tech setup on our streaming platform (Streamyard), preparing a FAQ that answers all their questions, asking them afterward if they had any feedback.

Scheduling the agenda

One of the biggest challenges turned out to schedule talks into thematic fitting blocks while keeping speaker’s timezones and scheduling conflicts in mind. I decided on six main blocks (e.g. “News & design”, “Datawrapper API”, and “Creative use of Datawrapper”). Five of my coworkers agreed to host one of these blocks – introduce the speakers and lead through the Q&A –, and they all did a wonderful job. Find the agenda here (PDF).

Promotion & communication with attendees

More than 1,400 people signed up to attend Unwrapped. Here’s what I did to promote the event and connect attendees:

• An article on the Datawrapper blog about every speaker, incl. a Q&A with them and showing some of their Datawrapper visualizations.
• Planning the conference website (thanks to my Datawrapper coworkers David and Jonathan for making it happen!)
• Posting about the conference on social media
• Reminding people who signed and informing them through nine newsletters, incl. reminders in the mornings of each conference day
• A pop-up Slack community that we opened in the week of the conference. It gave attendees the chance to ask questions. introduce themselves, and chat with other attendees.

Deciding on the tech stack & design

I took care of resarching and setting up the best solution for most of the technology involved, including sign-ups, newsletters, streaming platform (Streamyard), the pop-up Slack community, editing the videos, and more.

To help attendees to keep better track of when which talk will happen in their time zone, I set up a calendar with all talks that they could download or subscribe to.

I also designed the logo and, together with our head of design David Wendler, the overall look and feel of all conference material.

Nice things speakers have said about Unwrapped

Nice things attendees have said about Unwrapped

I posted this article elsewhere (most likely on the Datawrapper blog), which is why you don't see the full content here.

When should you use line charts and when small multiple line charts?