Evangelism Marketing: When customers do the job for you

The present day market is saturated. There are almost no marketing strategies left where you’ll not face any competition. Unless what you have is a groundbreaking product, it is very difficult to stand out from a crowded marketplace. Consequently, you need to keep looking for alternatives to sneak into people’s imaginations. Advertisements are a way, but how much can you squeeze into a half-a-minute long media file? Here is where the Evangelism Marketing comes into play.

Brand evangelism is a word-of-mouth marketing where customers voluntarily recommend your product to others, and will practically do the marketing on your behalf. People are less skeptical of a marketing evangelist as opposed to other hard-core selling techniques, simply because that person is not affiliated nor associated with a brand. As a result, there’s a higher chance of converting a potential lead into a definite sale when you use brand evangelism.

Evangelists can be carved out of loyal brand followers. If you are a startup looking to establish your presence in your chosen niche, you need to identify who your target market is even before you launch your product. Once the connection has been made, and the lead or prospect has been turned into a paying customer, you’ll face the challenge: turning that one-time customer into a customer for life. If you can even turn one loyal customer a month initially, it is a huge positive.

Establishing brand loyalty is important because these kinds of customers will not just buy your product; they also become your brand’s passionate advocates. You get new business based on word-of-mouth recommendations and referrals from them. The best thing about this? These loyal customers come across as credible, unbiased, and authoritative because you haven’t paid them a single cent.

This marketing strategy does sound quite lucrative and relatively simple. However, there are a lot of principles involved in it. The simpler the strategy, the harder it is to master it. So how do you perfect Evangelism Marketing so as to have an efficient group of loyalists doing your bidding? We’ll discuss these principles in the next post.

The five must-have elements of a data-driven culture

Many organisations claim to be data-driven. But are they, really?

To build a culture where data is understood and leveraged to its true potential, a business needs to address and solve these five key challenges.

Data glossary

As the saying goes, “it’s better to understand little, than to misunderstand a lot”.

Having your team on the same page, using the same vocabulary to discuss various data projects and metrics is incredibly important. It’s one of the aspects of a successful data culture that organisations tend to overlook. And therefore, one of the underlying reasons why arguments and misunderstandings between teams occur. A business should be proactive in drawing up an internal data dictionary that defines how an organisation understands various metrics. Without a clean and agreed-upon internal glossary, you’re almost always bound to have people interpreting data in their own subjective meaning-making process. And that’s where problems arise.

Image Credit: Konstantin

In most cases, the core metrics are understood equally well on all teams, and common definitions prevail. It’s the more complicated, deeper metrics that leave too much room for an interpretive spin that causes trouble. A successful construction of an internal data dictionary, and one that an entire organisation happily subscribes to, requires buy-in from all teams and management levels. In other words, it can’t be cranked out behind a closed door and handed down to teams to follow. The only way to invent a ‘data language’ within a business is to engage everyone. This might simply be the case of discussing which similar metrics could be collapsed into one common metric, fleshing out situations where deeper metrics would be more appropriate or talking through different business scenarios where data is used for decision-making.

In the simplest sense, making sure all employees are talking the same data language is one of the key building blocks of a strong data culture. You know what I mean?

Centralized, holistic source of data

Once you’re confident your organisation is fluent in data talk, the next major step is to establish a single source of data that feeds the whole business. This master data should be controlled and updated regularly to ensure the business is drawing from a high-quality, up-to-date information well. And most importantly, that all teams are working with the same data.

Image Credit: Diboune Juba

Having a centralized data source allows organisations to be more nimble and actually leverage the data they have in the day-to-day decision-making processes. If the data is hard to access or people don’t know it’s there, organisations run a risk of losing money or opportunities due to poorly informed decisions. Even something as simple as having a centralized data directory that points people to the best sources for certain data can make a big difference.

Data democratization

What’s the use of data if people can’t access it?

Data-driven companies are well-known for their inclusive approach to data sharing within the business. Of course, it doesn’t mean that everyone should have access to everything, but data democratisation holds immense potential for employee empowerment.

The most successful organisations tend to take great interest in assessing the data needs of all teams, not just the analysts and key decision makers. Even the customer-facing employees and back-of-house teams can benefit from increased access to valuable insights.

Image Credit: indro for Sebo

For instance, Sprig, a San Francisco based food delivery startup, gives their chefs access to an analytics platform where they can track and analyse all meals that have been ordered. This helps them better understand which foods, ingredients and flavours are popular amongst consumers, and which do not fare so well. Armed with this sort of information, chefs can quickly react to emerging trends and tailor their menu accordingly.

Essentially, it isn’t just about hooking your employees up with analytics dashboards. Instead, data democratisation helps organisations foster a culture where people feel comfortable handling and applying data in their day-to-day decision making. This can greatly boost any company’s performance and help cultivate new ideas.

Data literacy

Data literacy is the ability to derive meaning from data.

As soon as an organisation decides to follow the data democratisation path, it must critically assess its employees’ ability to examine and use data effectively. If staff can’t fully understand and apply the insights from dashboards, reports and analyses they encounter on a daily basis, the efforts to become a data-driven organisation are in vain.

Even the simple solutions available to all businesses, like training employees on data visualisation or basic statistics skills, could provide organisations with massive gains. Often, such seemingly unimportant decisions as choosing the right chart or the right colour palette can do a lot of damage if best practices aren’t followed. Inappropriate chart types and colours make data visualisations difficult to interpret, detracting from the value of data or worse, making the insights inaccessible.

Investing in continuous staff training to enhance data literacy throughout the organisation is a decision that will pay dividends.

Putting data to work

Building a data-driven culture is all fine and dandy, until those pretty-looking data visualisations and reports land on a desk of someone who’s ‘trusting their gut’.

Image Credit: Eran Mendel

Just to give you an idea of how big (and common) of a problem this is, Google’s Avinash Kaushik has coined the term “highest paid person’s opinion” (or HiPPO). These are the people with ‘years of experience’, who’ve seen and know it all, and they don’t care about your data insights if they clash with their very strong gut feeling. The Financial Times explains HiPPO this way:

“HiPPOs can be deadly for businesses because they base their decisions on ill-understood metrics at best, or on pure guesswork. With no intelligent tools to derive meaning from the full spectrum of customer interactions and evaluate the how, when, where and why behind actions, the HiPPO approach can be crippling for businesses.”

A data-driven culture needs to start at the top. Organisations that tend to value intuition over data insights are almost always led by people who don’t understand how data works or don’t know how to use it. But even that can be overcome with solid training and a data literate team.

Lastly, organisations don’t become data-driven overnight. Building a company culture where data is valued, understood and utilised is a multifaceted process, but once in place, it’s one that’s really hard to top.

The Art of Making Data Talk: Information-Roadmap (Part 2)

Top 3 Information-Roadmap Recommendations:

1. There is no ‘i’ in Team, meaning you need everyone. Subject matter experts (SMEs) should regularly gather together to continuously review processes and associated data-paths; confirming that defined goal(s) are continuously being met.

2. Report Cards. Expectations from this team of elected key-members would be to evaluate existing workflow processes, identify gaps and provide ongoing improvement recommendations.

3. Graduation/Storytelling. It’s important for everything to have a beginning, middle and an end. Providing staggered cycle-terms for team membership brings forward new ideas; allows for continuous innovation; keeps consistency and momentum going; sanctions knowledge sharing; crafts a comprehensive culture of openness; and upon term completion concur a sense of accomplishment.

Image sourced from: http://i.stack.imgur.com/LCGEZ.jpg

Related Posts:

The Art of Making Data Talk

The Art of Making Data Talk: Goals (Part 1)

The Art of Making Data Talk: Information-Roadmap (Part 2)

The Art of Making Data Talk: Communication (Part 3)

Data Analytics Use Case : BrewNation

Even for an early-stage startup, having good analytics is key to make good decisions and progress. BrewNation figured how to get fast and affordable analytics with the help of DataIntoResults. Let’s see how it works.

BrewNation is a marketplace for craft beer in France. They connect small brewers that make excellent craft beer to their customers. It’s a B2C business (they send beers at your doorstep) as well as a B2B business (they sell to bars and retailers).

BrewNation : The solution to bring craft beer from small breweries on your doorstep

Even being young, they have disparate systems and no easy way to analyse data in a meaningful way. The acquisition and website navigation is monitored by Google Analytics, the B2C sale process is managed by Prestashop and the B2B sales process use a custom spreadsheet.

Moreover, in order to have meaningful analytics, some reference data as well as some business rules should be used. For instance, a referral from a blog review (PR effort) is not analysed the same way that a referral from a brewery (supplier acquisition effort).

The system is based on modules (you can learn more about it in our manisfesto). The global overview of the analytics architecture is therefore the following. Each data source is ingested in the analytical database, blended and refined to ease reporting and insight generation. Google Sheet reports are feed directly by the database.

The Google Sheet dashboard is always up-to-date with fresh data. A cohesive view of the operations is only one click away.

Overview of the analytics stacks at BrewNation

DataIntoResults provide the data plumbing tools and the analytics database (PostgreSQL) to orchestrate all the business intelligence process every night. Some dashboards are powered by Google Sheet to report the activity and allow simple digging into the data. For more advanced research, the analytics database is only one SQL query away and fully managed. Moreover, as the data is already refined, queries are easier to write and consistent with the dashboard.

The following is the main sheet of the weekly report. Being a Google Sheet, it’s really easy to change the report if needed. For instance, the key metrics can change accordingly to the current focus of the team. No IT knowledge needed, it’s only a Google Sheet.

While the real dashboard is kind of the same, data are purely random.

When making a Google Sheet reports (or an Excel one), it’s good practice to segregate the data sheets (where the base data reside) from the reporting sheets. Therefore, you can easily change what you report without changing the data plumbing.

Example of a data sheet. Row 3 and below is automatically extracted from the database every day.

Having used the system for some months now, it is robust yet easy to evolve. It allows the company to be data-driven and be more focused on issues that can move the needle.

As a startup having reliable data is key to take decisions every day. Before using Dataintoresults we had multiple source of data which was time consuming for us. And as you know in a startup time is your most valuable asset! Thanks to Dataintoresults we have access to our data every morning in a very simple Google Spreadsheet. — Luca Fancello, BrewNation CEO

Are you ready to get more insightful analytics? Please contact us at contact@dataintoresults.com or you can use our tools by yourself. It’s free during the beta, so no reason to stay with low level analytics.

More Info: IDenTV’s Ad Tracking Portal

Hello and welcome to the free trial of IDenTV’s Real-time Advertisement Tracking Platform. Where we monitor over 60,000 unique Ads across the top National Channels by Ad-Spend.

Below you will find the An In Depth Tutorial & Monitored Channel List

ACCESSING OUR PORTAL — An In Depth Look to Maximizing Use!

To access our one of a kind real-time TV Advertisement Portal, please follow the steps below.

1. Go to https://portal.identv.com and read or reference the information below.

2. Sign up with a username and password (or with your social media account)

3. Be sure to check your email and click the activation link and read the brief instructions.

4. NEW ACCOUNTS ARE AUTOMATICALLY ASSIGNED AS A FREE BRAND TRACKING PLAN. Start typing the name of a brand and if we are tracking it will auto populate and refresh the REPORT.

If you wish to move to a paid service follow the instructions and billing. Select a plan that fits your needs. However, regardless anyone who signs up will by default have access to free brand tracking (up to 3 keywords/brands)! Done by selecting the “Free” plan.

5. In the “Select Ads” Section of the portal you can choose the ads and brands that you want to monitor and be sure to click update! (Start typing the name of a brand and if we are tracking it will auto populate and refresh the REPORT.)

6. To View Results Click on the “Reports” Section of the Portal to start receiving real-time data displaying reports on the selected ads or brands as they aired on TV. (Use the dropdown box to select a date range)

7. Exporting Data: Reports can be downloaded as a PDF or scheduled to be received via E-mail, by clicking the “gear/cog” icon in the top right of the “Reports” page (next to the “Edit” button).

8. “My Ads” Section of the Portal allows you to upload your own unique creative to be tracked if it is not already within our ever growing database of unique creatives! Be sure to enter the unique creative’s name or the official campaign’s name, upload and check back in a few hours for results by using the “Select Ads” Functionality described above to filter.

If you have any questions, need a customized package, or need additional support, please do hesitate to e-mail us at info@identv.com

Covering 120+ of the Top National Networks by Ad-Spend

5 ways % comparison lies and deceives you.

After working in the world of consultancy, I realized that when I switched over to business enterprises, clients are often misled by percentages. Statistical misrepresentation is your next door neighbor liar. It appears every so often, and people who use them are pretty sneaky conmen. Every time a big number pops up, your mouth involuntarily makes ‘ooh’ and ‘aah’ sounds, rendering you vulnerable to ‘alternative truths’.

I can still remember on one afternoon, my ex-colleague presented ‘engagement survey’ results to our clients, citing that employee engagement of the client is 10% below that of the industry average. However, what the heck does 10% even mean? Does it mean that one employee feels 10% not as engaged in their work as the rest of the industry? How does one person even feel 10%?!

Here are 7 ways, simply put, percentages have been deceiving you.

Trick #1: It makes your math brain dead.

Human deals stupidly with large numbers. We cannot conceptualize objects beyond our daily vocabulary (unless you are, of course, a mathematician). Sure, we can imagine ourselves rich as heck with numbers like 1,000,000,000 dollars, but we often find it increasingly difficult to count every single 100 dollar bills or do complicated math.

And so, using this fault in our perceptions, conmen trick you into thinking you’re getting more for less.

Do you remember the sales scheme your favorite store uses? Yeah, those 10% and then 30% off deals! When consumers are met with these seemingly cheap price tag, they are rendered dumb. Surprisingly, researchers at University of Miami, Rao and Haipeng Chen, found that majority of consumers treat percentages like numbers.

“…when consumers have to deal with more than one percentage at a time, they make errors that can be costly. For instance, if a store offers a 25 percent off sale with an additional 25 percent off for a certain product, people assume they are getting a 50 percent reduction. In reality, they are getting about a 43 percent discount.”

Your truly cheap discount may not be so cheap after all.

Trick #2: It blinds you from the real baseline.

I cannot lie that I have used this trickery before. Clients are astonished by really big percentages. Big numbers means it’s great right?! No.

Let’s take 447% as an example.

Have you noticed how this is something that has nothing to do with your sales team at all? Sure, if you use their software you may be better off than not having one, but does 447% still stand? Absolutely not, it’s a number that masks the baseline of percentages.

What is that 447% out of?

1 sales? If it is one sales, it just meant that using a software allowed you to sell 4.47 more donuts. Holy crap, now that is not as amazing as we previously imagined, let alone making our sales team happier.

Percentages also does not take into account the scale of growth. 447% may sound amazing for someone who runs a company of 10,000 sales per day, but is that truly likely? It’s quite absurd to imagine just one tool that can send sales to 4.47 million units of sales.

To really bring the point home, let’s imagine a bullet point a resume. Someone may claim that they have increased sales by 200% (or double the original amount). Since we do not know what the amount is out of, we may be impressed at first. However, if we found out that 200% growth means you only sold an extra donut from the beginning amount of one, it means shit.

Trick #3: It destroys variables in question.

Percentages destroy data integrity. It stops us from questioning whether the data collected is meaningful to what we’re measuring. Take being honest for an example.

Recall the independent fact-checking website of the 2016 presidential race that score how ‘truthful’ a candidate is within their speeches. According to their ‘analysis’, Donald Trump is only 4% truthful.

Wait what?

http://www.politifact.com/personalities/donald-trump/

What does it really mean? Does that mean for every 100 words Trump said he only said 4 truthful words? Which statements are we talking about? Nathan J. Robinson, author of ‘Why Politifact’s “True/False” percentages are meaningless’, hits a home run on this methodology.

That distinction makes an important difference. PolitiFact does not take a random sample of the sentences in Trump’s speeches. If it did, it would include things like “My father was in real estate” and “I’ve run a business for many years.” (Plus things like “You’re a beautiful crowd!”) Instead, it picks out certain statements that it thinks ought to be evaluated.

But, by their nature, these are going to be the most contentious and controversial statements in the speech.

Trick #4: It encourages misinformation.

Percentage requires context in order to make sense and we are often led to ‘assume’ the wrong context. Dr. Bruce D. Watson stated in his linkedin blog that criticizes an article by BBC New site that we are often misled by context.

We often see articles comparing company’s costs of doing something against its rivals. However, what we are missing is the scale in which we are arguing about. Take GDP growth of the United States and China for example. We often find debates on China’s GDP growth in the recent years as it has been on a downward trend. However, some advocates do say that China is better off due to its constantly large percentage when compared to other countries such as the United States as per the graph below.

GDP data provided by http://data.worldbank.org

From this graph, it may be evident that China is doing way better than the United States in general. China has been out performing in terms of GDP percentage growth. However, if we carefully examine the quantum amount instead of the percentage amount we will find that the United States, even though its percentage is way lower, looks way better off.

In 2014 and 2015, China’s GDP was 11,007.72 billion dollars and 10,482.37 billion dollars. While in 2014 and 2015, the United State’s GDP was billion 17,393.10 dollars and 18,036.65 billion dollars. Are your brains dead with the big numbers yet? Good, I just used trick #1 on you.

To help with visualization of data.

This clearly shows that the United States is better off, not only in terms of percentages, but in terms of growth and size of the economy as well. Don’t let yourself be misled by large percentages as graphs, try to find the real numbers and see for yourself what is going on.

I am not saying that China is doing absolutely horribly in terms of its economy, there are much larger things going on in China that, we in the western media, may not see.

Trick #5: It makes no sense.

Sometimes these tricks do not even need to be smart in order to deceive the average Joe. One can say that a percentage alone makes no sense, but take on warnings, several percentages together also make no sense. As I have said in the beginning about employees being engaged, what does the percentage truly means?

According to Aon Hewitt, a human capital consulting firm, 65% of employees globally are engaged. It’s 3 points more than last year! Wow, this is splendid. Wait, hold your horses. What does it truly mean?

Does it mean 6 out of 10 employees are engaged with their job? Does it mean employees feel 65% engaged most of the time? Does this number fluctuate over menstruation cycle of an average female employees?

The answer is: we truly don’t know. Sometimes these percentages are not insightful at all. Are we supposed to be worried that only over a half of employees are engaged? Are all employees supposed to be engaged?

At what % am I supposed to be worried?

Again, we truly do not know and at the same time it makes no sense. These percentages are turning emotions and feelings into math. We all know that emotions are not in anyway logical. We cannot control emotional percentages. So, don’t let these numbers fool you.

Take for instant, you’re 65% good at sex. Hell, that doesn’t even make sense. Fortunately, the consulting company took data on what makes employee ‘not engaged’ in their work.

Mission statements: act global, stay cool

…When Steve Jobs set up his own business, the genuine inventor came up with the idea of promotion. He claimed that only those products must be successful which meet the global needs of humanity. Steve Jobs expressed his desire to improve the current world with the technical inventions in the following way: ‘To make a contribution to the world by making tools for the mind that advance humankind’. This is what we call a mission of the company.

In case you are thinking about the business, you should not forget about the mission, as well.

Doubts and fears

Those who do not believe in the human power will call the process of mission creation as the waste of time. Time and time again I heard: ‘The only purpose we have is to make a profit. Why are you talking about such a pipe dream as the mission? We can change nothing except for our outcomes’.

Sounds cynical? Just one more detail: almost all the people who expressed similar thoughts failed in making business.

One more problem that acts as the brake on initiative is the fear. A lot of people are, actually, afraid of changes. They feel safe being concentrated on the promotion of the product without mission. They say, we haven’t done any bombshell, so, we are not responsible for the result which the buyers get.

We confess, it’s a failing practice which will never lead to a global success.

How to define the mission

Even though a good business plan and financial statements could help in the accounting, the mission is the fastest track to the customers.

Having understood how your potential target audience looks like, the next step is to answer several questions:

  • What are the needs of the target audience which the company is going to meet?
  • Why have those needs remained unsatisfied? What was the gap in the missions of similar companies?
  • Which level of service should be offered to the customers? Which approaches in the service do people generally prefer? How could your service differ from other companies?
  • What are the key values the company has?

Needs and values

As for the needs of the target audience, let’s look at the example of Coca Cola.

It was quite well-known in 1886 but it became an overnight success in 1920s. It happened because the advertisement matched the needs of people. The promoter found out: citizens of the country loved simple atmosphere, sports games and street clothing style. There was no point in presenting Coca Cola as the beverage of luxury or sophistication. The portrait of target audience described became the essence of the company mission:

‘The glass of fashion. Coca-Cola’s pure and wholesome refreshment is enjoyed by more people, of more ages, at more places, than any other drink’.

The next point is to talk about the core values of the company. They also should be connected with the needs of target audience. For example, the mission could rely on the ideas of luxury or try to stand for the concept of affordability. Perhaps, it deals with the sense of comfort, otherwise, it is likely to include the feeling of extreme motion. In all these examples the values differ, consequently, the mission statements would never be the same.

How is it to be famous?

We have collected some cool mission statements the international companies have. So, read and stay inspired!

– To inspire and nurture the human spirit — one person, one cup and one neighborhood at a time. (Starbucks)

– To inspire healthier communities by connecting people to real food. (Sweetgreen)

– To discover, develop and deliver innovating medicines that help patients prevail over serious diseases. (Bristol-Myers Squibb)

– To inspire humanity — both in the air and on the ground. (JetBlue)

– Build the best product, cause no unnecessary harm, use business to inspire and implement solutions to the environmental crisis. (Patagonia)

– ‘Provide children from birth through age 12, living in homeless or low-income situations, with the essential items they have to thrive — at home, at school and at play. (Cradles to Crayons)

Tableau in 10 Minutes: Step-by-Step Guide

Tableau is an innovative system of business intelligence enterprise-class, which can be used in both conventional and complex investigations: from visualization of questions and answers to complex data analysis (trends, correlations, and statistics). The convenience of the system is that it recognizes the data in any format, significantly reduces the time of data analysis by a visual presentation of information and is very easy to understand and use.

In this tutorial we will work with Tableau Public and Summer_Olympic_medallists_1896–2008 dataset. Tableau Public is free, so you can download it using just your email.

Let’s open Tableau Public. On the left side of the screen select the type of the data file and then open the Olympic medalists dataset.

You can preview all your data and choose only the needed data; simply drag and drop it to the field. Here you can change your data, split it or even join several fields by the columns with the same data. We will choose the first list of all the medalists.

After selecting the data, go to the first Worksheet. You can see all your data fields on the left side of the screen. Let’s drag the Discipline field and drop it to the Columns, then take Medal and drop it to the Rows. Click on the arrow near Medal (drop-down menu) and select Measure (Сount) -> Count to see bars that show the number of medals earned by each country. Another way to do it is to choose Number of records in the Measure tab and drop it to the Columns field.

The countries are sorted in the alphabetical order by default. You can sort them by the number of medals using the Sort icon near the vertical axis label or in the drop-down menu of Discipline:

Now we have a list of all sports disciplines, but there are such sports that had too little data (medals). It is not very interesting for us, so we can group them to one new field. Let’s choose these disciplines using SHIFT or CTRL and clicking “Group” after right-clicking the mouse.

Rename the new category to “Different” or “Other” or just exclude this data by mouse right click -> Exclude. Now, we have enough data for visualization and we can add some features. Let’s see the division of medals (gold, silver, bronze) in each sport. Drag Medal and drop it to the Color field on the left of the graph.

Now you can see that every column is divided into three parts. Use legend field Medal on the left of the chart to change colors of the columns’ parts and the sequence of the medals to personalize chart.

Also, you can add labels on your bars by clicking on Abc icon on the top (Show mark labels) or by dragging Number of records to the Label field on the left of the graph.

Note that all graphs are interactive. By clicking the mouse on the chart more information appears there. This allows us not to overload the original image. In this case, there is no need to add labels on such a diagram. You can see all the information just selecting the needed bar.

Now you can rename the Sheet by clicking 2 times on the name of the sheet. You can also add a new sheet by clicking plus (+) icon near the first Worksheet. Now we will show all the Countries in our dataset and create a beautiful interactive graph of the Most Successful Countries of All-Time. For that, we will need more information. Click “Add New Data Source” on the Toolbar and add our database ones more. Now drag the list TEAM EVENTS FIXED, ALL YRS TOTAL and go to the new Worksheet.

Let’s choose Show Me pallet in the top right corner of the window and select Country in the Dimensions field and Total in Measure (using CTRL). Show Me pallet highlights all kinds of graphs you can build using selected data. We will choose a Symbol map.

You can see in the bottom right corner of the scheme that there are 13 values with unknown geographical location. This is due to the fact that our base covers the period since 1896. We need to add all unknown locations to the already existing ones. Choose Map-Edit locations on the Toolbar. Now you have all 13 unknown countries on the top of the list. Change Unrecognized field to one of the existing countries. We will do so for every country except the Soviet Union. For this country we will enter new Location (it is optional, we chose 60 and 80).

You can change the size of the circles, their color, border and thickness using Size and Colortab. Also, we can add more information to this map and personalize it. Use Map –> Map Options on the Toolbar to change the map; add the Gold field (from Measures) to the Color tab field. Now we can change the color grade for the quantity of the gold medals in all countries and add to Detailfield all information you want to see when the country is selected.

Add a filter to a Filter field by dropping the Countries in it. Now you can choose only one country or a number of countries from the entire list to be shown.

Rename your sheet and add a new Dashboard, like it is shown on the image.

Drag and drop your Sheets to the central field. You can change Legend on the left side of your Dashboard. It is easy to change the size of this Dashboard, choosing its size from the drop-down menu. The laptop size is very useful to see the whole information without scrolling the page. Also, you can add Title or change the format of the Dashboard in the Dashboard menu on the Toolbar.

Finally, you can save the project to the cloud by clicking Toolbar –> File –> Save To Tableau Public. Now, you should set the name of the Book, log in to your account, and your Workbook will be available online. You can download the file on your computer, only after saving it in Tableau Public online.

Please note that all the Dashboards and Workbooks continue to be interactive online. You can share these graphs by copying the link under them.

Conclusion

In this tutorial we have shown only a small part of the Tableau functionality. This platform is a very convenient tool for creating various interactive graphs, maps, and charts, as well as their publication on the Internet. You can see all the possible types of data analysis of Olympic Games database here.

We hope this information was helpful and Tableau will be a useful tool in your work.

Originally published in datascience-school.com

An Introduction to Time-series Analysis Using Python and Pandas

So… what’s a time series and what makes it special?

From the initial data exploration, it was clear that we are dealing with what is known as a time series. Time series is just a fancy way of saying we are dealing with data points indexed in time order.

Usually, when dealing with time series we look for some special characteristics in our data to be able to make predictions based on it. Specifically, we look for a time series that is stationary.

Stationarity of a time series

We can say that a time series is stationary when its mean and variance are not a function of time (i.e., they are constant through time).

Stationarity is important because most of the statistical methods to perform analysis and forecasting work on the assumption that the statistical properties (mean, variance, correlation, etc.) of the series are constant in time.

How to test the stationarity of a time series?

Stationarity can be assessed in two ways:

  • Visually inspect the data points and check how the statistical properties vary in time.
  • Perform a Dickey-Fuller test.

Let us take a visual approach first and see how it goes.

By plotting the standard deviation and mean along with the original data points, we can see that both of them are somewhat constant in time. However, they seem to follow a cyclical behavior.

Although the visual approach can give us a clue, applying the Dicky-Fuller Test (DF-test) can provide a more precise way to measure the stationarity of our series.

Results of DF-test

I will not go through much detail on how the DF-test work, but let’s say all we need to care about is the numbers we see in “Test Statistic” and “Critical Values”. We always want the former to be less than the latter. And the lesser the value of Test Statistic the better.

Our series is stationary given that the Test Statistic is less than all the Critical Values, though not by much.

In case you ever need it, below goes the code I used to evaluate the stationarity.

def test_stationarity(timeseries):

# Determining rolling statistics
rolmean = timeseries.rolling(12).mean()
rolstd = timeseries.rolling(12).std()

# Plot rolling statistics:
orig = plt.plot(timeseries, color='blue',label='Original')
mean = plt.plot(rolmean, color='red', label='Rolling Mean')
std = plt.plot(rolstd, color='black', label = 'Rolling Std')
plt.legend(loc='best')
plt.title('Rolling Mean & Standard Deviation')
plt.show(block=False)

# Perform Dickey-Fuller test:
print ('Results of Dickey-Fuller Test:')
timeseries = timeseries.iloc[:,0].values
dftest = adfuller(timeseries, autolag='AIC')
dfoutput = pd.Series(dftest[0:4], index=['Test Statistic','p-value','#Lags Used','Number of Observations Used'])
for key,value in dftest[4].items():
dfoutput['Critical Value (%s)'%key] = value
print(dfoutput)

What if our time series was non-stationary?

There are some techniques one can apply to stationarise a time series. The two I am more familiar with are:

  • Transformation: apply transformation which penalizes higher values more than smaller values. These can be taking a log, square root, cube root, etc. This method helps in reducing the trend.
  • Differencing: take the difference of the observation at a particular instant with that at the previous point in time. This deals with both trend and seasonality, hence improving stationarity.

Pandas and numpy provide you with very practical ways to apply these techniques.

For the sake of demonstration, I will apply a log transformation to the dataframe.

# Transform the dataframe:
ts_log = np.log(data_df)
# Replace infs with NaN
ts_log.replace([np.inf, -np.inf], np.nan, inplace=True)
# Remove all the NaN values
ts_log.dropna(inplace=True)

Bonus track: We can even apply a smoothing technique over the transformed data set to remove the noise that may be present. A common smoothing technique is to subtract the Moving Average from the data set. This can be achieved as easy as:

# Get the moving average of the series
moving_avg = ts_log.rolling(12).mean() # 12 months
# Subtract the moving average of the log-transformed dataframe
ts_log_moving_avg_diff = ts_log - moving_avg
# Remove all the NaN values
ts_log_moving_avg_diff.dropna(inplace=True)
test_stationarity(ts_log_moving_avg_diff)

Clearly, we can see that applying log transformation + moving average smoothing to our original series resulted in a better series; in terms of stationarity.

To apply differencing, Pandas shift() function can be used. In this case, first order differencing was applied using the following code.

ts_log_diff = ts_log - ts_log.shift()
plt.plot(ts_log_diff)

Log-transformed data set after differencing

Let us perform a DF-test on this new resulting series.

ts_log_diff.dropna(inplace=True)
test_stationarity(ts_log_diff)

With the log transformation and differencing the test statistic is significantly smaller than the critical values, therefore this series is too more stationary than the original series.

Wrapping up…

When we face a predictive task that involves a time series, we need to analyze said series and determine whether it is stationary or not. To determine the stationarity, we can either plot the data and visually inspect the mean and other statistical properties or perform a Dickey-Fuller Test and look at the Test Statistic and Critical Values. In case the series happens to be non-stationary, we can apply techniques such as transformation or differencing to stationarise the series.

After all this analysis and preparation, the next step on the project was to forecast with the time series, but that’s a topic for another post 🙂

INVYO x Refinitiv, a data partnership to enrich financial institutions 🚀