If any teams are interested in a customized football data analytics solution, outsourcing their analytics effort or receiving bespoke data reports I've built a generic framework that I can customize for colour, logo, league, language and almost any feature. Below is a short demo:

I'm available for data analytics/visualization/science contract work if you know anybody who's looking for dashboard developers etc. Need not be football-related!.

Or newest public offering:

With Holoviz's Panel I can use Plotly click events to generate annotations on the fly. I can also add tooltips to graphs so that users know how to use them.

Using Holoviz Panel it's easy to save screen real estate via the Swipe widget, which is particularly handy when you have multiple graphs which don't look good when displayed in a small form factor.

I wish Heroku were even slightly more generous with their memory specs. Moving to Google Cloud has been immensely painful and unnecessarily complicated.

RT @SpreadexSport: Ahead of the Tyne-Wear derby we are giving away 5⃣ @SunderlandAFC home shirts!🔴⚪️. For the chance to win🏆. 1⃣ Follow us….

Here are Ivan Toney's stats from last season for comparison (I don't have Championship stats so it limited my predictive analysis). Toney does, however, have the advantage of having two years of EPL experience under his belt.

Should Arsenal be looking to sign Santiago Gimenez in the January transfer window? It's always hard to predict how a player would perform in a more difficult league, but the numbers indicate that he's worth serious consideration.

Hard to believe that, at 38, Gignac is still putting up stats like these.

RT @statiskicks: Huerta merece ser titular mañana.

Used @Panel_org to create a highly "reactive" web app. The built-in Tabulators, templates, and user authentication make the development process very easy and provide a good user experience.

I elected to limit my player search to players in England, Germany, Spain, Italy, France, Portugal, Holland, and Belgium. I also filtered for players between 21 and 31 who had recommendation match percentages greater than 70%. Here are the first two pages of matching players:

Thread: What if I wanted to find a replacement/backup for Erling Haaland? I used multiple machine learning algorithms that combined the VAEP framework, positional-specific criteria, and a recommendation engine to do so. First, here is a glimpse of Haaland's stats last season:

Turned out to be an SSL issue whereby the customer's local network was blocking insecure/non-SSL http connections.

Hey @MarcSkovMadsen and @PhilippJFR , have you ever seen a Panel app which displays fine for most users but appears blank for others as in this issue: