A data analyst who specializes in pandas has an answer.
It comes down to the user interface and how you design it.
“If you have a nice interface, then I would go with pandas,” says the man who calls himself the Panda of Data.
The first thing to understand about pandas is that it’s based on pandas’ data models.
“The only thing that differs from pandas and pandas-powered software is that you need to build the data models from scratch.”
You can get data from a website like Google, but the process is a lot less involved.
“Instead of building it from scratch, it’s just building a table with the columns you want to use and then importing the data,” he says.
The other big difference between pandas vs pandas powered software is the interface.
“With pandas you need a very sophisticated interface.
For example, you can add data to the table with a little bit of magic.
But with pandAS, you need the mouse to click the data and the mouse pointer to click on the data.”
It’s an issue for many people, but for Mr Cook, pandas was a necessity.
“It’s a bit like being a doctor who needs a better interface to perform an operation, so you have to use something that’s simpler,” he explains.
For Mr Cook and many other data analysts, the interface is crucial.
“As a data scientist, the main thing is getting the data right,” he adds.
“You want to be able to look at the data, use it and see the results.”
In other words, the pandas team wants to be the data analyst that gets things right.
But for many, this isn’t a realistic goal.
“I’ve seen it go from ‘what if you have millions of different datasets in a spreadsheet and you have no idea how to do that’ to ‘this is going to cost me a lot of money to build a nice UI for a dataset’,” Mr Cook says.
But the data is worth it, Mr Cook believes.
“Data is a very powerful tool.
I’ve had a few instances where I’ve made a mistake in the way I’ve built a dataset, but I’ve never been able to see it.”
He has built an example of the data in question that shows how he did it.
It’s a dataset that is highly dependent on other data.
“For example, in a dataset where I have a lot more data than the dataset that I have, I’m going to have a bigger problem than if I just have a single dataset,” he notes.
“In this case, you’re going to get more errors if you’re trying to build an app from scratch because you’re building the app from the ground up, instead of building a model that you’ve got to go back and modify a couple times to get the data that you want.”
For data analysts who are looking to build their own software, there are some things that are a bit different to other data analysis software.
“We want to build software that’s as fast as possible.
For me, the most important thing to build is a fast, robust and fast-performing data analysis suite,” Mr Cook explains.
Data analytics software has been around for a while, and there are a few tools that are still around that are more suited to data analytics.
One of these is StatViz, which was created by John Gruber, who has been at Twitter for nearly 20 years.
“There’s a lot we can learn from StatVid,” Mr Gruber says.
“StatViz was really an attempt to make a data science tool that was fast, scalable and fast to use.
It didn’t have the scale of a database, but it did have the speed of a spreadsheet.”
It’s easy to use, fast, and fast for data analysts.
“But there’s a bunch of other things you can do.
For instance, you don’t need to spend as much time on data analysis, because it’s the same data you’ve been doing for a long time,” he continues.
The StatVist tool lets you build your own custom data analysis engine.
“Basically, you write a function, you run it on a dataset and you get results that you can use in your work.
That’s really what it does.”
StatVist is also free, but that doesn’t mean that it won’t cost you money.
Mr Grubers says that he is looking at alternatives to StatVis and that he has some ideas on how to get it free.
“In the short term, I would like to offer the free version of StatVizer, which is basically a tool to automate the process of generating data and analyzing it.
That would be free for anybody to use.”
There are also tools that come with StatVista that help you work with data