Today we will see how to create simple plot graphs in Python using Seaborn. I have found in a Data Science book (by Sinan Ozdemir) a simple graph where we can plot sales and expenditure in advertising for different media like TV, Radio, Newspaper.
Let’see how it works. Where is better to put the money? First of all import pandas and seaborn packages and dataset
import pandas as pd
import seaborn as sns
data = pd.read_csv('http://www-bcf.usc.edu/~gareth/ISL/Advertising.csv', index_col=0) #import data in CSV format using Pandas data.head() # let'see how data are structured
Let’s plot the data using some magic… and using seaborne package we can say to Python, to plot data having 3 x_vars, based on our 3 first column of our database, and on y_vars sales. Let’s see the result
But image to write the same commands now with Radio & Newspaper with capital letter. Python is very sensitive (or at least Anaconda version that I’m using), so take care of using correct name of variable.
Where to find correct Variable name? You can see it from:
the head of your table
In Anaconda, on the Variable explorer in column Value
If you need further info on magic function, %matplotlib inline, you can have a look to this post on Stackoverflow
Did you like this post on a simple plot graph. Share it to friends, colleagues or martians using below buttons and do not forget to subscribe to our fun newsletter.
An image worth more than thousand of data, sorry words. Yes, sometimes you have a lot of data to present in a very small period of data. What is better than a good dashboard where to see in a glance your KPI’s or a nice infographics to show tons of info in few seconds?
People don’t have time, so having all in one page is very useful to present but also to make a good storytelling to help your audience to digest complex info and memorize important messages.
What now you have to do is just think about which are the data that you would like to present and how to create a good storytelling that your audience will remind.
Dashboard helps you to understand immediately what is going well (maybe showing green numbers, up arrows) and where to investigate more maybe with other self-service reports.
An example of interactive dashboard -Photo by rawpixel on Unsplash
To create powerful visualization you need to fulfill the following requirements:
– What I want to explain with this dashboard? Maybe I want to show if we have reach our sales target, or which are the most contributors for growth or products that are in delay
– Test how simple and easy to read is your dashboard: go to one of your colleague with less familiarity with technology and ask to explain the content of our report. If he/she report the right message you have created a good one. Otherwise interview other people on what is difficult to read or unclear and simplify.
–Create your dashboard: you have several tools to create it:
Excel: Best info At Chandoo.org where you will discover how to create and manage your dashboard.
Python: More complicated but you can define every aspect of your dashboard.
Plotly and Bokeh are the modules that you can use to excel on this topic.
An interesting example is this Bokeh dashboard or Kickstarter project by category and status (successfull, cancelled…) including also name of the project if you pass through
An example of kickstarter dashboard done in Bokeh
R: Best choice: Here you can customize everything, using Shiny and Rmarkdown using less code than Python.
An interesting example is R Cran download monitor, where in one page you can see evolution of package download, name
CRAN download – R dashboard example
If you want more please consider in subscribing to our fun newsletter and our recent posts
Today I would like to discuss with you about statistic skills you need to develop, if you want to become a data scientist. Do you need to be graduated in Statistics to do this job? The quick reply is some Statistics is needed, but practice is more important.
Data Science without Statistics is like owning a Ferrari without brakes. You can enjoy sitting in Ferrari, show off your newly owned car to others, but you can’t enjoy the drive for long because you would crash land soon!
-use statistic skills to explore and visualize data
-most important statistical theories (like hypothesis testing, bayesian analysis)
-know most common statistical models and define which is best to use (like linear regressions. Time series analysis)
-evaluate if the model is working for the purpose of your analysis.
But theory is not all, so the best way to learn about Stat skills is through practical approach. So don’t expect to become a good data scientist only reading books or learning theory.
Use statistics skills to explore data:
Understand & summarize your data:
If you are new in the world of data, dataset and graph, you can start from this free course : Analyzing categorical data provided by Khan Academy. Here you will learn how to identify individuals, variables, read different types of graphs and much more. I suggest to stop at first module, if you are at a basic level.
Grouping & Visualization
This is a fondamental exercise to be done with your dataset. Let’use a free dataset on regarding Wine reviews Kaggle dataset.
In this dataset you will have 130k wine reviews with wines coming from all over the world, scored by wine taster from 0 to 100, including a lot of info related to qualitative features of the wine.
We would like to understand in few minutes:
which variable do we have in the dataset?
which data types are (numbers? if so what kind of numbers – maybe strings of text)
How many data we have? Number of rows and columns
For some variables which are minimum, average and maximum values
which is the median (mid point of the data set) and the mode (most frequent observation)?
How is the distribution of our data? For on?e variable which is the average observation for the first 25% of my dataset, and for 50%
Let’s briefly report same simple statistical concept that it will be deep dive in separate post
Descriptive statistics: you are probably familiar with mean, median, mode, ranges and quartile. This info will help you to understand how looks like your dataset.
Coming back to our Wine dataset just with one command you can identify many of these information. In this case you will see that your database has around 130.000 records, with an average points (coming from reviews) of 88,45 and a reported average price of 35,36$
Simple statistic skills: descpritive statistics with pyton and pandas
Minimum value is 80 and 4€ for price and max is 100 for variable points and 3.300 for price (Wow!!)
Percentiles:25%, also called first quartile: it means that observation 32.492 is represe
nting 25% of your dataset (in ascending order). This observation has an average review of 86 and a price of 17$.
Interesting to see that to arrive to 50% of this database you will increment only 2 value in points (88) but +30% in price (25%)
Distributions: explain you how it is possible (probable) that your data will be distributed. More famous is normal distribution, also knowned as “bell curve” (that happens many time in nature). Another important distribution curve, is binomial, that easily represent two status, i.e success or failure of a new drug.We will discuss about distributions in a separate post about distributions.
In the next topic we will discuss also about Hypothesis testing, Regression model, Time series analysis and other Intermediate Statistical concepts
Stay connected and subscribe to our newsletter to learn more about how to became a great data scientist.
If you have liked this post on fundamentals statistic skills for a great data scientist, please sharing it through social buttons. Let me know your comment or thought adding a comment!
Ready to learn how to analyze data with Python in few minutes, without knowing too much about Python language? In this brief Python tutorial, you will learn how easy is importing 130.000 rows in few seconds, with Pandasmodule for Python. With less than 10 line of code you can explore number of records, column names, and start to know more your data with some exploratory analysis like mean, max & minimum.
But let’s go, step by steps. What is Pandas?
Pandas eating “your data”
pandas is an open source library providing high-performance, easy-to-use data structures and data analysis tools for the Python programming language.
Let’see the code to import CSV. Less than a minute
#Import module pandas
import pandas as pd
#Using Panda to load csv
Location = r'C:\DATASET\WINE REVIEWS\winemag-data-130k-v2.csv'
df = pd.read_csv(Location) #Read CSV in location
Attention: change your path, changing ‘C:\DATASET\WINE REVIEWS\winemag-data-130k-v2.csv’ with your path. If you are not familiar with this check it out: 3 simple way to change your path in Anaconda/Python
What you will learn at the end of this Python tutorial on Analyze data in 5 minutes using Pandas
Undertand variables of your dataset with df.dtypes
Main statistics with df.describe
Find missing value
Group info with .groupby, i.e by country
Filter/subset your data
Export your data in Excel
Learn other useful command like df.head/tail
WHAT KIND OF VARIABLE WE HAVE IN THE DATASET?
In the first part of this Python tutorial, we will use Anaconda and Pandas to start exploring our data. We can see that now our df (Dataframe) has around 130k records and 14 variables(129971,14)
Using command:
df.dtypes
you can easily check what kind of data do you have. Here we have basically all variable as an object , while first variable, points are integer number (int64) and price include floating number (with decimals)
If you want to learn more about which are most common Statistical & Math skills for a data scientist, discover our dedicated page.
PYTHON TUTORIAL: LEARN MAIN STATISTICS WITH ONE COMMAND:DF.DESCRIBE
Let’s focus only only on numerical variables.
Point: This represent a number from 0 to 100 as scoring of wine taster.
Price, no need to explain
Here we introduce a great command, df.describe, to analyze mean, standard deviation, min, max and quartile distribution of our data:
df.describe
We can see that variable Points are available for all data (129.971) and has a minimum of 80 and a max of 100, with an average of 88,45
Our first 25% of dataset (32.492) has an average point of 86, 50% of dataset (88)
Of course you can do the same in Excel, but you need to create several cell and write several formulas. So now you have learned in this Python tutorial how to save some time in exploring data.
PYTHON TUTORIAL: UNDERSTAND IF YOU HAVE MISSING VALUES IN YOUR DATASET
Another powerful command to analyze data with python is understanding the quality of your dataset. Do you have missing value? How many? In which variables?
Let’s continue our Python tutorial, where soon you willsee the value of using Python instead of Excel.
Write these two lines of code and you will find how many missing values you have in your dataset for every columns
WHICH IS THE AVERAGE SCORING AND PRICE IN ANY COUNTRY?
One of the most common things to analyze data with Python, is understanding average data, maybe grouping for some of your variables.If you want to know which is the average score and price by country, you can use
.groupby ().operations,
where in the parenthesis you need to put variable to be grouped and after the operation that you want to do .mean or .sum for example
df.groupby(‘country’).mean()
group by
So we will discover that in our dataset, average price in Argentina is 24,5$ with an average of 86,7$, better than Austria that has 90 point in average but you have to pay 30$
Of course you can groupby by multiple column (‘country’,’region’)
FILTER OR USE SUBSET WITH WINE WITH PRICE >90$
Maybe you are interesting to easily know how many permutations you have in your database that fit with a particular threshold. In this case we would like to know how many wines has a price >90$. Result is more than 4.000 records!!!
df[df.price>90]
HOW TO EXPORT IN CSV OR EXCEL WHEN YOU ANALYZE DATA WITH PYTHON ?
Easy, just write
df.to_csv('first.csv') #creating a csv file called first
df.to_excel('first.xlsx', sheet_name='Sheet1') #creating a xlsx file called first, in sheet 1
OTHER USEFUL COMMAND TO ANALYZE DATA WITH PYTHON
Df.head= See head of your dataset
Df.columns= show your columns names
Df.tail[3] = show latest 3 record of your dataset
Df.index= show you the range of your dataset
Stay tuned next time we will describe how to user wordcloud to describe string variables. Subscribe to our newsletter for more news
Don’t forget to leave your comment and to share the article if you like it.
Today we will discover more about data mining. If you are not familiar with this concept, it is better that you start to understand more what is behind. We are talking about powerful tools and techique that will help you to get insight from your big data.
A simple definition is:
Data mining is the sum of technique and methodologies to collect information from different sources and manage in automatic way through algorithms and logical patterns
How data mining could help you to collect data ?
Data are growing fastly not only on social and open database, but everywhere.
With data mining tecnique like data scraping (taking data from internet, like ecommerce price , weather data, stock exchange…), you can increase number of datasources that you can use for your analysis.
Did you know that you can get data also from images. Discover more here:
In few minutes with very small line of code you can learn how to web scraping data using Python and R
How to group your data: clustering analysis
Image a big databases with many customers. It often happen that you have a lot of different groups of customer . Clustering analysis could easily identify which are customer with affinity that you can address in a similar group target, maybe because they are similar to size order, purchase need, purchase attitude.
Example of cluster from www.dummies.com
This will help you or your firm to set different pricing, product and general marketing strategies more focused for that particular target.
Using Python or R will help you to identify clusters (see below an example of 3 clusters)
Regression analysis: identify future output based on historical data
Consider a dataset with icecream sales of last three years and one with temperature information. With regression you can create an algorithm to estimate how much icecream you can sales based on expected temperature
Interesting article that clarify more regression, expecially on marketing
Anomaly detection
Yes, how many times you have seen dataset with errors like typo distraction or duplicated info. Through specific tools and Machine learning you can easily identify and prevent this kind of error analyzing historical data and suggesting correct value.
How much time you can save from more robust and clear data set? Data scientist usually pass from 70% to 90% cleaning data
Classification analysis: a powerful data mining technique
In this field are growing machine learning algorithm and chatbot that in future could try to solve most of our questions, maybe about a product features, classifying our question base on common patterns.
Could be also interesting to identify common words in books, text, maybe through Wordcloud.
Signup to our newsletter to know soon how to analyze through wordcloud any text with Python and discover more info on datamining tools and techniques
I have started this website with a clear mission: to help all people that manage data, like Data analyst, Financial Controller, IT/IS people, Business partners and who else in the business that would like to base their decisions starting from robust data and analysis.
After many years as Finance Controller, Business Partner and Pricing Manager I have seen that people tend to maintain status quo. So changing is difficult, what I have discovered during my experience is that changes are more easy if you have robust data behind that help you to understand reason to change and benefit of new solutions.
Following Fromzerotodatascience.com you will:
– save time in your analysis using tips and tricks and most powerful tools (i.e Excel, Python,R)
Data scientist, what an appealing definition, but what kind of skills and which tools do you need to start? What is a good definition of data scientist?
Let’s discover on the web some good resources that will help us
Bernand Marr, in his “9 steps to become a data scientist from scratch” makes a good syntesis. I just reported what I like most:
Math & Statistical skills: mmh, not very appealing in some cases. I remind my study in Statistics and how much theoretical looks like. Probably I will discover soon that could be used in a more pratical way.
Learn tools: most of the activity will be cleaning your data to make good analysis. Remember garbage in, garbage out
Community: life of data scientist in same cases is not easy, many different tools and languages to be used. Having some peers and community places where asking help and support is fundamental.
Practice: someone has said that 90% of what you will learn is training by doing. I don’t know if the percentage is realistic, but certainly what I have learn most is from my trials, success and fortunately from my mistakes
4 easy step to easy installing Python with no admin rights
Ready to install Python. Yes, but if I would like it to have it in portable way (maybe on an external hard disk?). You soon discover how you can manage this.
temperature is rising, holiday are coming. Yes we are in Summer. Why don’t get this fantastic time to learn more about Data Science? Today I will introduce to a good practice called SoDS
The first way to became a data scientist is to start to doing. Why not starting with Summer of Data science (SoDS)?
#SoDS 2018 was created by Renee @BecomingDataSci with the goal to learn something new about data science. So, now you are wondering how to partecipate.
Is very easy:
In Week 1, you just start thinking what you want to learn. Just make a short list of things maybe regarding different topics (maybe have a look to my article “ How looks like a good data scientist”)
Maybe are you interested to learn more about math & statistical skills or you want to know more about Tools like, R, Sas, Python, or maybe find mates for a data competition on Kaggle
So are you ready for SoDS Week 1?
You have one week to search on the web and decide what you you want to learn.
Write in the comments what do you like to learn. It will an helpful way to decide to really start to become Data Scientist and to have a lot of fun.
I’m Frank, and I would like to share with you my journey on Data scientist. Discover with me what we need to learn to become a successful Data scientist and what is behind this world.
I’m not an IT guy, but I like helping people to take decision based on numbers.
How many business or personal decision was taken with wrong data or no data at all?
So follow me and we will learn together tools and techniques to treat informations, present it and make the difference in your personal and business environment.
What I want to do?
I would like to make a simple synthesis of resources that are available on the net and see how we can use it. Through this exercise we will learn to be more productive and efficient in analyzing data.
What you have to expect?
Dummy & simple language, limiting at minimum technical and abstract concept and lot of interesting analysis.
See you soon for my first topic!
Write in the comment any topic that you are interested to focus, I will do my best to reply