Building A Financial Dashboard In Python With Dash - Part 2/3

Huge respect for the video! Amazing project structure. That's what I was looking for! I worked several years with R/Shiny and now extending knowledge to Python and this videos is an amazing start indeed!

Hello Arjan. First I'd like to say you are a true asset to the Python community and refer to you as Santa PAUSE. Virtually every thing you speak to is a nugget of wisdom and knowledge that necessitates me to pause and think. There is always a link between the object and the path to get there. Using very simple models has been so effective for me as I try to wrestle with how data is flowing from component to component or module to module. My advice to your audience is to be prepared to think (pause) if you really want to absorb your message. Thank you sir.

Really good stuff Arjan, I set myself on the journey of creating more financial insights and wanting to learn more including dash. So this is really helpful. And still enough to explore on your channel. I'd like to get your thoughts on the following in your video. In the data schema, there is a definition of CATEGORY = "category". In updating the values in the bar chart you use:
filtered_data = data.query(
"year in @years and month in @months and category in @categories"
)
If you would change the category name in the data schema, you would also have to change the filtering query. Wouldn't this defeat the purpose of defining the data schema?

P.S. by changing it, it would become not really readable anymore (although there might be a more pythonic way)..
"year in @years and month in @months and `{0}` in @categories".format(DataSchema.CATEGORY)

Your content is really great, practical with focus on software development really like it.

I've been away from programming for a while and the world has changed and YOU very efficiently explain how to implement the new ideas. I had always wished for the concepts you present to come into being, especially data and presentation separation. Watching work is like plugging in to the matrix and getting a download.

Is there a trick to get the refresh to work, without re-starting the application?

The training was excellent.
How can I display the information stored in a Mango database instead of displaying the graph and modify and save each record?

My OCD is screaming at the red scripts at the top of editor.

man please add part 3

Login for Dash via Azure or/and Google? Please?

Isnt using streamit much faster ?

@ArjanCodes Could you please make a video on default parameters for functions in Python? I came from C# and the idea that those parameters are bound and persist between executions is weird. Maybe there is some idea behind this behavior?

Great content man !!

Raised the bar on this one, Arjan. 😁

Would be great to see integration of a multi paged dash app with django

Love your Channel! Please, when you got time, i would appreciate so much if you deep into Logging module, especially the logger hierarchy and how to set logger for different scripts which are related.

Thank you so much!

Great contents! One small suggestion: you can use the natsort package to sort the months correctly.

len() of a dataframe is also number of rows.
So
if not len(df):
Is the same as
if df.shape[0] == 0:

Very refreshing to see this series. I am already working a long time with Dash, and it is very nice to see another devs approach in creating Dash apps.

One suggestion from my side, in the latest versions of Dash it is unnecessary to group outputs, states and inputs in lists in the app.callback decorator. I would even strongly argue against its usage. If you are not careful in how you handle function input parameters and return statements, you can run into weird bugs, as the list formatting might require you to format your function inputs and outputs as lists as well. If you just supply the outputs, inputs and states as separate input parameters, you save yourself potentials headaches.

You're a life saver !!

Love this but would be much more excited about a frontend module like PySimpleGUI. It is much simpler to write and has more widgets, each of which are more customisable

It was very interesting, but what I found most valuable was the Schema pattern!

I literally went back right now to my current open pull request to add this 😁

Really cool and concise guide! The only thing I'd suggest is that pandas can define columns as categorical data, which might be useful for your last column. I don't think it would have any impact on the current code, but it might allow some useful methodology if it were extended

I am also playing around with Dash and Plotly but I do run into a wall when I want to connect it to code which is based on asyncio. Looks like the problem is that Dash is based on Flask and Flask does not like async calls. There are alternatives to replace Flask with Quart but for me this a huge struggle and pip projects doing this are beta at most.

Maybe you can do a part 4 where you connect a separate backend to Dash. Either in the same codebase (my usecase is ib_insync as dependency where I am building stuff) or even doing some REST calls to a completely separate backend.

After watching this series I’m scared to walk into my kitchen at night and find a red squiggly line staring at me

After watching this series I’m scared to walk into my kitchen at night and find a red squiggly line staring at me

@Arjancodes it's like you are reading our minds. Just got tasked to do budgeting for our research grant, this is so timely! Can't wait to dive in

Great content as always I recently build in node its been nice to see your approach. I find this funny or at least intriguing... With covid and recent economic events in the US Prorgammers started day trading the volatility like no other. The idea of value is changing and people are building systems like this to manage their problems. Not sure if you have access to Plaid but their baseline requirements to use their API has dropped massively maybe due to more volume.

Great I have on medium also some tutorials on that topic