Market Landscape: Segment Lab

Segment Lab

The Segment Lab is the heart of the Market Landscape. You use this module to create segments which in turn help you to look at the data that is relevant to you.
A segment is a set of terms that can be used across the whole platform to search through the Scitodate data. The search engine of Market Landscape can search for these keywords in the title, abstract, body and keywords of scientific papers. Alternatively, the terms can be applied to search in the data Scitodate has on grants, such as the title or abstract of the grant.
Therefore two types of segments exist: Paper and Grant segments.
Paper segments can be applied to papers, authors and institutes
Grant segments are only applicable to grants

Creating Segments

You can create a segment by going to the Segment Lab, by clicking on the "Segments" button at the top of the screen. From there, follow these steps:
Click on the "Add new segment" button
Enter a name for the segment. We advise to add a descriptive name that you can later use to find the segment.
Select a type from the dropdown menu next to the name input field. Two options exist, paper and grant.
Set a name for the first block. We advise to add a descriptive name that you can later use to understand the context of a matched item.
Add at least one keyword. Try a term you are familiar with.
Click "Fetch Results" on the right top of the screen and review the results.
Click on "Save Segment" next to the segment type.
Congrats, you have successfully created your first segment! 🎉

Building Segments: The Basics

Building a segment can be hard and needs practice. Don't let that be a discouragement; You can iterate on your segments as often as you want.
It is a good strategy to revisit segments you are using repeatedly to apply your newly found knowledge.

Blocks

Segments are made from blocks. Blocks are a collections of keywords with an operator and search criteria.

You can add a keyword to a block by entering a term you want to search at the location indicated by step 5 in the previous section. In other query builder platforms you have to define logical rules between keywords in order to make a proper search. In the Segment Lab, we have re-invented the search query and made in more visual. This should make it much easier for you and your team to understand the query and its results..

You can add multiple blocks to a segment. In fact, you are most likely to add multiple blocks in each segment you create! Adding a block is simple, click the "Add block" as indicated in the image below:


The next step is to give your block a name. This name is not going to affect your search query. However, it does help you to understand the structure of your segment. This is essential when you revisit the segment at a later point in time, or share the segment with a team member.

For example, you are looking for the people who are working on an application areas. We can create a block for specifying the keywords in relation to the application area and another block for scientific techniques which are used in that area. For this, you can create one block called Application Area and another called Techniques. You can create a virtually unlimited amount of block, e.g. Devices, Company Names, etc.

If want to use some block separately from other blocks, it might be better to split these into separate segments.


Blocks are not save automatically. you should save them by saving the segment.

Now you have made your block and gave it a name, it is time to write add keywords to the block. Every time you add a keyword, another keyword row is automatically created. You can add a virtually unlimited amount of keywords to a block.

"OR", "AND" and "NOT"

You might have already noticed the buttons for "OR", "AND" and "NOT"; this is an essential element of the block.
Use OR if any of the terms should be found in a paper or grant.
Use AND if all of the terms should be found in the paper or grant.
Use NOT if none of the terms should be found in the paper or grant.

Let's give an example to illustrate how this is useful. Say you are creating a segment in the field of cell biology. The first term you have added naturally is cell biology. However, you notice that this does not cover all the papers that you want to retrieve. You also want to include papers matching biological science. It does not matter to you whether the paper matches cell biology or biological science, so you choose an OR block.

Next you want to drill down into some details. You are interested in cell biology papers that regard protein visualisation, for this you create a separate block. Rather quickly, you realise that you are particularly searching for papers that describe protein visualisation in the context of viruses. Thus you add the term virus and set the block to use the AND relationship. Only papers that both mention protein visualisation and virus will show up in the results. Note that the relationship between blocks is of type AND, thus the papers in this example must also match cell biology OR biological science.

Lastly, you notice that some unwanted papers as they are in the field of plant biology. You now add another block, this time a NOT block, and add the term plant biology. Voila, you have effectively used all three types of blocks.

Block settings

You might have already noticed the block settings, which has many knobs and buttons. It may look a little intimidating, but it is not as complicated as it looks!



Field
With the block settings, you can control what parts of a paper or grant should be considered for search. If the checkbox before the "title" is checked, the title will be included in the search.
Papers have the following fields:
- Title
- Abstract
- Body
- Keywords

Grant have these fields:
- Title
- Abstract
- Keywords

You can also configure how important you deem the field to be. Perhaps a keyword matching in the title is of much greater importance to you, as the paper may be much more relevant. The field importance impacts the scoring of items, items with a higher score appear higher in the results if you sort by this metric.
There are four options for importance:
- High
- Medium
- Low
- None

When checking the box for No Stemming, keywords will be restricted in terms of spelling flexibility. In other words, the search engine will look for the exact term in results. The option is unchecked by default. With No Stemming disabled, the search engine will look for derivationally related forms connected to the base of the term such as common conjugations, plurals etc.
Some examples:
- "_microscopy_" will not match to "_microscope_" with No Stemming enabled
- "_microscope_" will match "_microscopes_" with No Stemming disabled

Minimum should match determines how many of the keywords should be found in a grant or paper for the item to match the query. If set to 0, no terms have to match, but the system will show them if they do match.

The Slop defines how many words may occur between words in a keyword. Consider the case where we have a slop of 3 and the term "cell biology". For this example, "cell" and "biology" may be separated in the paper by two words. The phrase "stem cells methods in molecular biology" will match the term "cell biology". When the slop is set to 0, the term has to match exactly.

Term settings

You might want to set block settings for only one term in the block. In this case you can use Term Settings, which can be found to the right of the term input field.



The checkbox indicates No Stemming for the specific term.
The number input field determines the slop for the specific term.
If you click on the "disable" button, the term is deactivated and will not be used in the search.
You can delete the term by clicking on the red "X" button.

Segment Preview

You can already preview search results in the Segment Lab. This preview is optimized for building a good segment, not for exploring data. We recommend to use this preview for the following use cases:
Evaluation of the segment quality. E.g. by checking if known researchers that should match the segment, actually show up
Exploration of terms. Find new related terms by looking at the context of currently matching terms
Evaluation of terms. See how terms are matching by looking into their context. This can prevent any unwanted papers or grants from matching unknowingly.

Building a good segment is an art that requires practice and we highly recommend training by our customer success team.

Using Segments

Segments can be used in all other modules, such as the Market Feed, the Prospector and the Analyzer. The segments will typically be used as filter: Any document that matches the segment, will show up as a result.
It can also be used as an exclusion filter: Any document that matches the segment, will not show up as a result.

Sharing Segments

You can share segments within your company from the Segment Lab by clicking the "share" button. Here you can select a set of users or a team that should have access to the segment. An additional message, describing why you are sharing the segment, can be included. You can also copy a link to the segment if you want to share it via a specific channel, for example an internal messaging tool.

There are different scenarios why you might want to share a segment. For example, your marketing specialist may request you to create segment as you have the right expertise in a specific field. Segments created by experts of a field are generally of a much higher quality.
You might also want to share the segment for reviewing the quality. There are many situations where you are part of a multi-step process, such as market analysis and prospecting. The market analyst and the marketing specialist might want to reuse the same segment for different purposes.

Updated on: 27/07/2022