Visualizations combinations

I want to visualize electric load profile weekly wise. The electric load profile has the electric load readings taken for every hour on each day.

Typical CSV columns are date,day,hour,load.

In a day 24 readings are taken, in a week (7 multiplied by 24)=168 and in a month (168 multiplied by 4)=672 are taken. So there are around 700 rows for a typical month.

I want to have visualizations for weekly wise, day wise, hour range wise etc., As of now I am just using only two columns with serial number and load. And two visualizations I made.

One as x axis serial number from 1 to 700. And y axis as load.

Second as x axis load and y axis as count.

How to have different combinations of visualizations for day wise and hour wise.

I am using python client to create elasticsearch object and using CSV reader to read the file and populate elasticsearch object.

Please guide me. thanks,

What does your data structure look like?

Hi Warkolm,
The data structure looks like below.

Date,day,hour,load in kW
1st April 2019,Monday,1 AM,50
1st April 2019, monday,2 AM,72
.
.
1st April 2019,Monday,9 AM,200
.
.
1st April 2019,Monday,11pm,55

2nd April 2019,Tuesday,1 Am,45
.
.
.

31st April 2019, Wednesday,1 AM,57
.
.
..

And what does it look like in Elasticsearch?

Since I don't know how to show date and day in visualization I made CSV with two columns serial number,load because they are numbers. I know only to create visualization with numbers.

I created one histogram with x axis as serial nbr and y axis as load.

Second visualization I made with count as y axis and x axis as load. So I can see one particular load how many times occurred in a month.

How are you loading this data into Elasticsearch?

You should really be converting the date, day and hour into a single timestamp field.

Ok Warkolm. Then I have to find how to write python client code for creating timestamp and populate that into elasticsearch object. So you mean to say I will have only two columns ? Timestamp, load.
Is my understanding correct ? What is the visualization then for this kind of data structure ?

Ideally you would have those, yes, plus the source of the data (which I guess is a meter?).

That way you can do changes over time using line/bar graphs and separate by source.

Sorry, I didn't understand. So should I have data structure like this ?

Date, day, time, Timestamp, load in kilowatts.

Could you point me to some documentation explaining visualizations having Timestamp and also data having non number data.

Like this;

{
  "timestamp" : value,
  "reading": value,
  "source": value
}

If this is in CSV, then you can easily ingest it using Beats or Logstash, without having to write your own code from scratch.

Condemn my less intelligence in elasticsearch ... :frowning:

I feel I have to work more on understanding beats and logstash and how to interpret data structure in relation to elasticsearch.

By the bye, what do you mean by "source" ? I understood "reading" which is electrical load. But what is the "source" which you are mentioning

No worries!

Regarding source; I am guessing that you have load readings from different sources?

Thank you warkolm.

The "load" value taken from an electric meter hourly basis manually by from the meter display. This is the single source for me.

Ok. I was just thinking you may want to be able to differentiate if you end up with multiple different sources, but sounds like it's not relevant.

Yeah. Multiple sources of data as of now not there. But good that you have given valuable insight what if there are multiple sources of data. I will have to explain this.

Could you point me to some visualization involving timestamp. I am having little difficulty in understanding how a timestamp actually appears in a typical visualization. Because I always assume visualization means only numbers on x axis and y axis. Eager to know how timestamp appears in a visualization. Thanks

https://www.elastic.co/guide/en/kibana/current/most-frequent.html for eg.

Also take a look at our blog, there are tonnes of examples from ourselves and our users.

Thanks warkolm for your mini tutorial in this post. I will go through the blog and the examples.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.