This is how we built a football dream team for under €100M
#Moneyball
We modelled 14620 active players using
@EASPORTSFIFA
data to study the skills of valuable players. Matching these skills with undervalued players, we picked out the hidden gems💎(1/6)
We turned this dataset from
@Renfe
with all the train trips in Spain into this beautiful graph - map.
Take a look at the shortest path connecting 2 of the most dense and populated regions, Andalucía and Murcia by train being so close geographically :P
We found from this Spanish designer salary poll :
- Designers from BCN makes significantly less and are less happy than their peers from Madrid
- The gender pay gap exits but junior women are making more than men
- Making more than 25K makes the difference
Dado que el tema de salarios 💰 es un tema tabú, he creado este Excel abierto para que podamos escribir, de forma ANÓNIMA, qué cobramos y cómo trabajamos.
¿Te animas? 👇
Me gustaría hacer un hilazo para recopilar, conocer y visibilizar las COMUNIDADES de mujeres en tecnología que tenemos.
Diseño, UX, IU, web, software, datos, devops...DE TODO.
Me ayudas?
(Contesta a este tweet, porfis).
Gracias.
We looked up the number of tweets/minute mentioning "earthquake" and "Granada" (Spain). The seismograph registered the earthquake at 12:15:25, just after 17 secs, hundreds of people started reporting it on Twitter. These tweets actually traveled faster than the seismic wave! 🤯
Almost 15K people wrote a tweet saying what are their favorite programming languages, most hated, which one they would recommend to learn programming... We collected and analyzed with
@graphext
all the answers to the meme. Here is a thread with the main insights we found 👇
Because at
@graphext
we are data people who love design 💞 and
@mendesaltaren
are design people who love data... we rented an office together in Madrid, in calle Preciados 7 😀
Mbappé was supposed to replace
@Cristiano
in
@realmadrid
. But, is he really the best fit? Analyzing with
@graphext
71 variables based on their skills and performance we can see both players are quite different.
We analyzed with
@graphext
all 2018 posts on Reddit
@slashML
to see what topics were more engaging among people interested in Machine Learning this year and this is the result
After the Catalan elections, the number of tweets containing insults against pro-Spain Catalans or against Catalan separatists has increased significantly, especially the most offensive ones.
Where are the cities in Spain with at least one company registered every 12 people vs where is the unemployment rate higher? with
@graphext
+
@populate_
data
There are many amazing women working with data in our industry and in academia. We made this Twitter list including hundreds of them and connected them on this graph
#8M2021
Check out the interactive viz made with
@graphext
👇📊
🛠 Graphext Hack - Web Scraping
Yep, you can collect and analyze data from web pages directly in Graphext ... 😲
Choose 🗒️Text - Topics and point to your URL column 🌐
Graphext will scrape the page content and automatically apply NLP - extracting useful language features 🧰
We got a good sample of data from unstructured images and text to learn how to look awesome on video calls, analyzing 1K screenshots from
@RateMySkypeRoom
ratings of people appearing on TV from home during the lockdown ->
La cuenta de
@OT_Oficial
ha dado bastante más visibilidad en la última semana a Amaia, Alfred y Aitana, mientras que a Miriam y especialmente a Ana Guerra se les ha promocionado mucho menos de lo normal
#OTFinal
#OTDirecto5F
Over 25% of
@JulianAssange
's tweets (430) since September were about Catalonia. His total activity grew exponentially since the referendum, like a never before, not even during the US elections going against Hillary
In the first episode of the final season of
@SiliconHBO
, Richard could have used
@graphext
to cluster all those messages to classify them in an unsupervised way instead of building his own tool!
#SiliconValleyHBO
We made this interactive
@graphext
map of the Spanish data analytics community
📘 The community around
@databeers
events and academics
📙 Data journalists like
@AlbertoCairo
@cabralens
or
@kikollan
📗More business and marketing people, consultants
Thanks
@TwitterMktgES
for inviting us to participate in a panel discussion about
#TwitterxContent
. We talked about how
@Newtral
uses
@graphext
to analyze the key political opinions in Spain and visualize how different communities interact with one another during their TV shows :)
You can now go and play in
@El_Pais
with this
@graphext
interactive embedding and see how the brand new Spanish Parliament is connected and polarized on Twitter
At
@graphext
we use WebAssembly to be able to create our fast and interactive data analytics app on any browser...
@miwelc
and
@crispamares
are talking about our experience using it at the C++ users Meetup in Madrid today
We bought a copy of
@fisherdanyel
and
@miriah_meyer
new book "Making Data Visual" after listening this interview in
@datastories
. It talks about the process of moving from vague data questions to building actual visualizations that solve real problems
📖 Docs | Getting Started
We've created guides and walkthroughs to help people grow their skills.
• Datasets 🧮
• Projects 📂
• The Graph 🕸️
• Compare ⚖️
• Trends 📈
• Insights 💡
It just got much faster to become an expert user.
Diversity and the gender gap has been more discussed by VCs than AI or crypto this year! These are the most shared links by the top 1000 VCs in the US talking about
@JamesADamore
memo,
@sacca
,
@davemcclure
or
@caldbeckj
Ever wondered how pop songs are changing?
We used topic detection & clustering to pick out the themes of 5100
#Billboard100
song lyrics | 1964 -2015
Then we analyzed their - sentiment & emotion - with
@Cardiff_NLP
|
@huggingface
NLP models
💕🎵
(1/5)
This thread summarizes pretty well what's wrong with Business Intelligence == Dashboards. At
@graphext
we believe dashboards are ok to monitor some KPIs, but you can't understand why and how to change these metrics from looking at isolated line charts! 📈
The problem is that it's usually not possible to generate meaningful insight simply by looking at line charts in a dashboard, regardless of how much interactivity the analyst jams in. Even if it were, people simply don't want to spend their time clicking around on dashboards.
Analyzing with
@graphext
this dataset containing records of every historical figure with Wikipedia biography in 25+ languages
we can see some interesting patterns when we connect them by simmilarity 👇
Introduction to NLP for Text Analysis |
#DataAcademy
"NLP is a marriage between linguistics and mathematical modelling. 🗣️🧮
Text analytics uses NLP to study human language in such a way that enterprises can learn about what people think and feel"
Another
@graphext
insight: because of the Oscars we have to wait until the end of the year to watch all the good movies!
82% of the movies that won the academy award for best picture in the last 17 years were released in the last 3 months of the year.
We used Image Analysis 📸 &
@GoogleCloudTech
Vision API to reverse engineer 40000 Movie Posters 🎬 from
@IMDb
1. Upload IMDB data
2. Connect CloudVision API
3. Run image analysis
4. CloudVision deconstructed poster features
5. Graphext clustered posters
We collected every tweet in 2020 from 38 UK news organisations to find out what the media have been reporting on.
Then we visualised categories of tweets as trends to see what the British media landscape looked like throughout the year👇
wrap it up, folks! I just proved that hot dogs are, unequivocally, sandwiches 🌭
embeddings for different foods from the
@OpenAI
API, reduced dimensionality with UMAP, plotted here colored by "category". Hot dogs are solidly in the sandwich cluster, along with tacos and burritos
We used the
@Apple
data to cluster with
@graphext
countries that are experiencing similar behaviour of changes in people walking outside:
🟠Nordic, Germany slowly going back to the streets.
🔴Italy, Spain and other suddenly decrease by more than 90%
🟢Slowly reducing since Feb
Our friends from
@seedtag
built this impressive Graph exploring
#Euro2020
players and teams. ⚽️ 🏆
They linked players to teams, teams to groups and groups to the cup to create a hierarchy of entities in the tournament.
A seriously cool
#dataviz
of European
#football
.
Why use Graphext for Exploratory Data Analysis in 2023, even if you already master Pandas & Python.
@victorianoi
replicated
@Rob_Mulla
analysis visually much faster, more intuitive, and powerful, like understanding the correlations between variables.
🛠 Graphext Hack - Plotting Maps
Yep, that's right. You can use Graphext to build
#maps
.
To start plotting flow or connections between places, choose a Geospatial analysis type.
#data
#DataVisualization
#hack
"For doing survey analysis, I recommend Graphext, it is a great tool for data analysis, that help us to have insights very quick, without coding a line of R"
@Dmartincc
from
@zensei_app
about how he used
@graphext
to understand their users personas
@graphext
looks like a very very cool project, with a generous free tier, for the ones who want to get insights on the benefits of graph knowledge encoding. . Graph AI is the next AI revolution. Access to context is absolutely essential for intelligence.
At
@graphext
, we are making the impossible possible: working with datasets of millions of rows in a browser.
We are looking for programmers who know C++ and/or TypeScript to continue building on top of this impossible front
Managing portfolios is a tricky business 💸
We spoke to
@Parga_Fran
and his team at about how they use Graphext to analyze the composition of financial portfolios 🗂️
Usually takes - 2/3 weeks
With Graphext - 2 days
(1/4)
Cristiano vs Messi - Madrid vs Barcelona for startups - the Popular Party vs Ciudadanos . Comparing the differences and similarities of complex things made out of hundreds of attributes is now super fast and easy with
@graphext
:D
@Aquaservice
use prediction models to calculate the number of water bottles to load into each of their delivery trucks every day 🚚
To spot patterns in their model's error (improve-iterate), their data science team cluster delivery routes w Graphext! 1/3
💫 We added a new function in
@graphext
that will merge similar categories with similar spellings. Real word datasets are plenty of variables where you will find duplicated values that were misspelled or written in a different way. We used char2vec under the hood to implement it
Just released - Logarithmic Plots 💎 You can now unlock hidden insights with our new logarithmic scale in data visualizations - it will help you see your data in a whole new light.
#DataVisualization
#LogarithmicPlots
Analyzing with
@graphext
millions of orders from
@Instacart
to understand how the 42K products they offer are bought together. So many options to choose candy and ice-cream! :P
Analysing the connections among the 650 characters that appears together in scenes of Friends with
@graphext
:
- Rachel & Ross are clearly the main characters, they even have their own community.
- Joey is the most detached from the group.
- Chandler is the least important.
🚀 Plot | Special Product Update 🚀
We're extremely excited to announce the release of our new analysis panel: Plot 📊🥳
With Plot, create bar charts, heat maps, box plots and all of the time-series visualisations previously found inside Trends! 1/4
It's worth reflecting on why this might be the case. I think often when we generate value from data, it's a result of a *dialog* we have with it. As opposed to knowing the exact question in advance and setting some precise process in motion to answer it.
We made this integration in less than 30 minutes with our social media data to create amazing automatic reports with
@fivetran
+ BigQuery +
@graphext
, it updates automatically with one click 📊😍
Looking at hundreds of variables to predict who will win the Oscars. Here are some interesting insights we found with
@graphext
: longer films tend to win more Oscars
It took more than 10 days for
#Tabarnia
to appear in the national and international news today. Another viral that needed slow cook on Twitter before it reached your grandma...
Thank you so much to the over 100 people that came to our first customer event yesterday! We record all the talks and will publish them as soon as we have them :)
Exploratory Data Analysis |
#GraphextDataAcademy
"The analyst firstly aims to become familiar with the contents of a dataset, spotting anomalies and understanding value distribution. Then they can begin to transform the data so that it is ready to model"
We are looking for someone to help our clients make the best out of our product and translate clients' suggestions into new features and refinements of the platform :)
More details about the position and apply!👇
Data Podcasts 🎧 are as varied as the inputs to a Linear Regression model 📈if you know where to look 🧐
We've curated a list of 35 of the best🏅:
📚 Real World Stories
👨💼 Business Analytics
🧪 AI & ML
👩🏫 Tutorials
📊 Data Vis
... and more (1/4)
🛠 Graphext Hack - Exporting Data
When you build a project, Graphext will often transform and enrich your dataset.
Head to the details panel to export the data and carry on your analysis elsewhere with the useful variables that were added.
You know how every app trying to get you on a chat 🤖? Well, at Graphext, we decided to do the complete opposite 🛑
Introducing tell-you-what-you-need-to-know data analysis. Dive into your metrics, and BOOM 💥 Graphext tells what else to explore...
#ByeByeChatbots
📢 Today we officially launch TweetNLP, an all-round NLP platform for social media. From sentiment analysis to emoji prediction and more 🔥🔥
✔️ TweetNLP includes a Python API, a demo and tutorials. Useful for developers and researchers alike. Want to know more? 🧵
These are the 4 Americas that emerge when we connect all the counties by the amount of americans that moved between them,
@graphext
automatically clusters the most connected regions. The bigger the node the more people left. Cook county (Chicago) had the largest population loss
"There's a gap between what I do and what I think I do."
Here's how we hacked private data for a social experiment with
@Ballantines_ES
.
The project exposes how our digital behaviours often contradict the way we perceive ourselves.
Transforma tu negocio con Data Science, el 7 de Mayo empresas como
@bbva
,
@OgilvyES
,
@TwitterEspana
y
@interiorgob
nos contarán como ya lo han conseguido en un desayuno organizado por Graphext.
Reserva aquí tu plaza:
Over the moon 🦘🌙 to be featured in the
@moderndatastack
newsletter alongside a bunch of other talented people & data companies!
In case you didn't know already - we're really passionate about the data ecosystem 👩🔬👨🔬
2/ Featured tools this week
-
@graphext
is a no-code data analytics tool for the modern data stack. More powerful than dashboards and more intuitive than notebooks. It helps you to build data science projects remarkably fast, collaboratively, and without writing code.
Exploring Customer Churn: sharing one of our favourite datasets:
- 7,043 rows, 32 features
- Clustered customers so you can spot patterns related to churn
- Explore it with Graphext or export the data to play on your own
🔗 Interested -
#dataset
#eda