When you create a project you build a recipe to transform your data. From the moment you pick a type of analysis through to the moment you execute your project, Graphext combines a customised set of algorithms to process your data. This is a crucial part of your analysis and determines the output of your project.

Recipes are organised into types of analysis. Choose a type of analysis that matches the content of your data. Recipes come in many shapes and sizes meaning that the process of building each one is slightly different.

‍

General Instructions

Recipes come in many shapes and sizes meaning that the process of building each one is slightly different. These instructions outline the general process of building a recipe using the project setup wizard.

‍

How to Build a Recipe - Generally?

1. Start from the datasets panel of your team's dashboard.
2. Select a dataset to start working with it.
3. Pick a type of analysis from the left sidebar.
4. Inside the window belonging to that type of analysis, select an option that further matches the type of analysis you will conduct.
5. Complete the set of steps that appear in your left sidebar. These are specific to individual recipes. For details on how to follow these steps - see the recipe step instructions listed below.
6. Once you have completed the steps, click 'Continue'.
7. Enter a name for your project.
8. Select 'Execute'.
9. Done ... Now Graphext will build your recipe!

‍

Step by Step

The combination of steps involved in building a recipe will depend on the type of recipe you are building. Follow the instructions on each step here.

‍

Data Enrichment

Data enrichment increases the amount of information in your dataset. Graphext will use your existing fields to extract additional details and add these into your project. This might involve extracting precise date information such as months, days and times from date fields.

For a complete list of how you can enrich your data, see the data enrichment index.

‍

How to Enrich Your Data?

1. Start building a recipe using the project setup wizard.
2. Pick a type of analysis.
3. Choose an option from the sidebar to further indicate the type of analysis you want to conduct.
4. Inside the recipe wizard belonging to the type of analysis you have chosen, open the 'Data Enrichment' form.
5. Select the 'Enrich your data' dropdown.
6. Choose an item or items from the dropdown.
7. Complete the forms that appear below to configure your data enrichment.
8. Done ... Your enrichment selections will be processed once you execute your project!

‍

Aggregation

When building a recipe you can aggregate data to combine points that share features into collections. Aggregation will summarise your data points around a variable.

Note that not all analysis types support aggregation.

‍

How to Aggregate Your Data?

1. Start building a recipe using the project setup wizard.
2. Pick a type of analysis that supports aggregation.
3. Choose an option from the sidebar to further indicate the type of analysis you want to conduct. The title should include 'Aggregate'.
4. Inside the recipe wizard belonging to the type of analysis you have chosen, open the 'Aggregation' form.
5. Select the dropdown menu inside the 'Aggregation form.
6. Choose a variable or variables to aggregate your data.
7. To delete a variable from the aggregation list, click the 'x' icon next to the variable name.
8. Done ... Your data will be aggregated when you execute the project!

‍

Network Visualization

Setting up your network visualisation defines how your nodes will appear within your projects Graph. Customizing the appearance of your Graph makes it easier to recognise patterns and understand their meaning.

As you build a recipe the setup wizard will prompt you to specify how your network visualisation will appear. Different types of analysis feature different options for customising network visualisations.

‍

How to Configure Network Visualizations?

1. Inside the recipe wizard, open the 'Network Visualisation' form - sometimes will be hidden inside 'Advanced Settings'.
2. For each of the dropdown menus inside your 'Network Visualisation' form, choose a variable from the menu list. This associates a variable with the focus of that dropdown. For example, you can choose a variable to represent the size of the nodes in your Graph.
3. Done ... You can change these settings later!

‍

Data Extraction

When building a recipe you can extract new variables from text fields in your data. Occurrences of keywords, emojis and adjectives among other items can be pulled out of your text. Extracted items appear as new columns in your data meaning that you can later filter points according to whether these items are featured or not.

Data extraction refers to extracting key values from text fields in your data and is only possible for recipes focusing on text analysis.

‍

Analysis Types Supporting Data Extraction

Text

Surveys

Commerce

Google Analytics

Social Media

‍

What Can Be Extracted from Text

Significant Terms

Sentiment

Emojis

Keywords

Mentions

Hashtags

‍

‍

How to Extract Data From Text Fields?

1. Start building a recipe using the project setup wizard.
2. Pick a type of analysis that supports data extraction.
3. Choose an option from the sidebar to further indicate the type of analysis you want to conduct. This should be related to text analysis.
4. Inside the recipe wizard belonging to the type of analysis you have chosen, open the 'Data Extraction' form.
5. Choose the information you want to extract from your text using the dropdown inside of the 'Data Extraction' form.
6. To delete an item from the extraction list, click the 'x' icon next to the item.
7. Done ... This information will be extracted from your text when you execute the project!

‍

Clusters and Network Creation

Clusters are groups of similar rows in your data. They consist of groups of nodes that are linked based on their similarity together inside of your project. When you build a recipe analysing the relationships between points in your data, each point is linked to other similar points. The resulting connections form your project's network, which is visualised in the Graph.

Creating clusters helps you to zoom in on patterns within your data. You can use factors to create networks and clusters in most types of analysis.

‍

Factors are variables that will be considered when creating links in your data.

Targets are variables that are key performance indicators. They are the variables that you want to gain a deeper understanding of.

‍

How to Create Clusters and Networks?

1. Start building a recipe using the project setup wizard.
2. Pick a type of analysis.
3. Choose an option from the sidebar to further indicate the type of analysis you want to conduct.
4. Inside the recipe wizard belonging to the type of analysis you have chosen, find the 'Clusters and Network Creation' card.
5. Start by adding a factor.
6. Find a factor to add from the list of 'Other Variables' on the right side of the card.
7. Select the checkbox next to the name of the variable.
8. Click 'Send Here' on the first box under the 'Factors' column.
9. Add more factors by repeating steps 6-8.
10. To add a target, find your target variable in the list of 'Other Variables' on the right side of the card.
11. Select the checkbox next to the name of the variable.
12. Click 'Send Here' on the first box under the 'Target' column.
13. Done ... These variables will form networks and clusters from your data.

‍