Deepnote = Jupyter notebook for collaboration 🤝 | Deepnote Tutorial

Key Takeaways

data scientists don't work alone most of the time they work in a team which means they need better tools for collaboration let's say you are a data scientist you are building a jupiter notebook but your teammate has a question on a specific cell maybe they want to execute it and give comments on that what if i tell you that it is possible to have google docs type of feature in jupyter notebook where someone can comment on on a cell they can execute you can even publish jupyter notebook as an app when you add google docs type of collaboration features to jupyter notebook plus 10 other cool features what you get is deep note deep node is basically a jupyter notebook on steroids all right enough talk let me straight away show you a demo here is the website for deep note you can sign in for free i'm going to use my google credentials unless i am working in some finance company in a team called equity so i'm going to create the workspace called equity and i can invite my team members here on this screen you can connect to a data source by clicking on this button here you will see list of popular databases with which you can connect for example i want to connect with mysql you can give hostname port your credentials and you'll be able to pull data from that my sequel database directly into deep note i will straight away go to my workspace now this is how my home page looks i'm going to show you some pre-built notebooks so here i will go to finance dashboard now click on daily stock price dashboard now here this is just like a jupyter notebook but it has some additional cool features so i'm going to just show this code here we're using yahoo finance library for time series and i'm just going to execute this and then execute this particular cell as well and it is showing you a stock price for microsoft i it is an interactive chart so you can already see the visualization is better compared to what you get in jupyter notebook or similar and you can then run the entire notebook on a cloud see here this is your cloud configuration just like google collab now your teammate can comment on on your sales so for example let's say someone wants to see the prices of tesla they will go here your colleagues will commend that can i can i see [Music] can i see tesla here so you can have just like google docs you can have interactive uh conversation here you can see all your comments clc i just commented and then based on that you can make changes so i'm gonna change this to let's say tesla run this and the chart is changed to reflect the price of tesla you can also share and publish so when you click on share you get this unique link which you can open and you can see that same notebook all right now let me show you how to publish this as this as an app so when i select open publishing editor here on the left hand side i have a jupiter notebook on the right hand side it just looks like an application a web application where i can hide all my cells for the code and i can just have the visualization okay or the ui elements of that app and when i say publish changes hooray it got published see all the celebration when you open it you realize this looks like general website you know a web application where you can interact with the chart this is more for uh non-technical users you know who are the end consumers of your notebook in this editor you can hide certain cells you can move them around you see all this control so you you can control the anatomy of that ui application so this is so cool that you wrote a jupyter notebook as a data scientist but you don't have to go to software engineers and say hey can you build this proper software app you can just directly convert this notebook into a working application let's go over the right panel here you have integration which means you can connect with different data sources now they have given a postgrade database for website analytics and if you look at view schema see they have one table with all these columns this is just a sample integration you can connect with a different data source by clicking on create new and you will see variety of data sources let's say you want to connect with amazon s3 or redship you can click here provide all your credentials and you will be able to connect you see all the comments here can i say tesla then you also see the history kind of the history of the execution of your code when you run a notebook it is obviously running in a cloud just like google collab you can also schedule a notebook so let's say you want to run this notebook daily at 5 pm or something like that then you can set that up when the execution fails you can get notification by email as well and this is pretty cool let me now show you a different notebook so i'm going to go to user analytics and click on user retention charts where you know for a website they are showing you some user retention analysis now here they already have the sql cell so let me just create a new one so if you click on a block here you get different options and am going to select sql and in the sql it is already connected with website because in my integration i have the integration with the postgre database of the website analytics or data set that they have given me i can write an sql query here so i'm gonna write the same sql query by the way and if you see the schema see i'm selecting few columns from this table and when you run it the results will be saved in df underscore 2 which is my pandas data frame so my code got executed and i got my result as a pandas data frame now here i can do sorting you can also do filters just like excel so you get excel kind of feature in jupyter notebook here in deep note actually and you realize that this is pretty good for the people who are not very technical people who don't know pandas and python too much in detail now they are working in this environment where some of this is code but some of this is a user pure user interface for example you want to filter on something this is pure user interface okay so this is short of like low code uh development environment i mean there is obviously some code but compared to jupiter notebook this is low code where you can do some of these things using uh ui you can also add visualization so i'm going to click on add and just say chart and you know df2 data frame now here i can plot different things i think this doesn't make sense maybe let me show you uh this one a b testing because that has a better better data set for the visualization and you can see the conversion rate chart here uh for your a b testing so on our website you have variant one and variant a and b you are doing a b testing of a website and you can see a different conversion charts directly in this notebook i really enjoy working with deep note and i encourage you to check it out [Music] you