FAQ

Question 1

If I have a web widget and I change the data behind, what happens?

Answer 1

A widget is like a photograph. When you save a widget, you save both the visualization settings (color, icons, filters) and data (nodes, relationships and positions). The data is not automatically refreshed based on the database content (but can be manually refreshed).
For example, if you delete a node from the DB this node will still appear in the widget when you open it.

Question 2

What happens if several people work on the same data?

Answer 2

LKE manages the concurrent edition.
If two people edit the same visualization, a pop-up menu will appear on the browser of the last user doing an action on the same visualization saying someone else is already working on it. This user will then be asked if he/she either wants to take over and override the other person’s work or not and continue in read only mode.

Question 3

What query languages do you support?

Answer 3

We support Cypher (Neo4j) and Gremlin (JanusGraph, CosmosDB)

Question 4

Can queries add/delete/edit into nodes and edges?

Answer 4

If the user has read/write rights for queries, yes

Question 5

Can you edit 20 nodes at the same time directly via query?

Answer 5

Yes, generally speaking you can perform the same queries available in the graph DB

Question 6

What languages do you support?

Answer 6

Today LKE is only available in English. Adding new languages is part of our roadmap. This is something that could be fast tracked if it is a hard requirement for a project and if the client is willing to help us fund this.

Question 7

Do you support geospatial visualization? How does geospatial visualization work? Can I use my own geospatial data?

Answer 7

Yes, it’s possible to plot on a map the nodes that have lat/long coordinates. See here: link
By default, we integrate various geospatial services such as Mapbox or Open Street Map. You can switch to your own geospatial server if it makes sense for security reasons.

Question 8

Can you use the geo-mode offline?

Answer 8

By default Linkurious Enterprise uses an external server for the map, so it works only online. However, by changing the configuration, you can use an offline map. Please refer to this link. The map is composed of .png files called “tiles”. You would just need to put these tiles in an offline source that it’s reachable by Linkurious Enterprise server.

Question 9

Does Linkurious Enterprise support APOC?

Answer 9

Linkurious Enterprise supports APOC but your APOC query:
- MUST have a return statement
- MUST return a graph, nodes, edges
- CANNOT return virtual graphs
For example you can run the following query:
MATCH (a:AF_NODE)
SET a.sha1 = apoc.util.sha1([a.label])
RETURN a

Question 10

Do you support database X?

Answer 10

If we don’t support database X:
focus on pitching Ogma which can be connected to virtually any backend
evaluate if the client is committed to database X, if he is not it may make sense to suggest using LKE and switching to one of the dbs we support
if the client is not interested in Ogma, explain that we’d consider supporting database X if a client is ready to sponsor such an integration

Question 11

Are you compatible with RDF databases?

Answer 11

The off-the-shelf Linkurious Enterprise is not compatible with RDF databases at the moment. In order to use Linkurious Enterprise, you’d need to sync your data between an RDF backend and a backend compatible with Linkurious Enterprise (Neo4j for example). See here: link

Question 12

Is Linkurious Enterprise compatible with any version of Neo4J?

Answer 12

Yes, we support all recent versions of Neo4J (both Neo4J community edition and Neo4j enterprise).

Question 13

How does Linkurious Enterprise integrate with Neo4j?

Answer 13

You can use one of the Neo4j connectors: http, https or Bolt. Further details: link

Question 14

Are there any additional considerations when the data in graph db is updated constantly?

Answer 14

This depends on your indexing strategy. If you want to constantly update data, we suggest using neo4j-to-elasticsearch or neo4j search which can automatically update the index as soon as the db has been modified. More details here: link
Otherwise the updates done on the data without using Linkurious Enterprise will not show up in the search index until you manually reindex the data.

Question 15

Do you support a cluster of Linkurious Enterprise containers (for performance and scalability)?

Answer 15

Currently we don’t support clustering for scalability, we support it for fault tolerance. Please see here in the doc how to achieve this scenario.

Question 16

How to integrate Linkurious Enterprise with machine learning?

Answer 16

Linkurious Enterprise is a visualization and analysis tool used to find insights in graph data.

Our clients are using ML to:
create the graph: ML can be used to identify entities and relationships within unstructured data
detect anomalies in the graph: ML can be used to identify suspicious patterns such as fraudulent transactions
automate the review of alerts: ML can be used to analyse the manual review of alerts and automate the process

You can see an example of integrating Linkurious Enterprise and ML here: https://linkurio.us/blog/cgi-detect-fraud-prevent-financial-risk/

Question 17

How to integrate Linkurious Enterprise with graph analytics?

Answer 17

In Linkurious Enterprise there are a few ways you can leverage graph analytics via the visualization styling, the query templates and the alerts:
visualization styling: you can map the size or color of a node based on a graph metric (eg the PageRank score)
query templates: you can launch within the Linkurious Enterprise UI a custom graph query (eg finding the shortest path between two nodes, identifying communities)
alerts: you can detect graph patterns via the alerts dashboard (eg identifying groups of clients interconnected via personal information, identifying clients with a PageRank over a certain threshold)
When it comes to computing graph metrics or identifying communities across a large graph, we recommend data scientists to run these workloads with dedicated tools (eg the Neo4j graph algorithms library when using Neo4j).
You can find more information here: https://linkurio.us/blog/graph-analytics-linkurious-enterprise/

Question 18

Is there a way we can get multiple filters added to Linkurious for end users to select / filter nodes, edges or properties.

Answer 18

Yes, this can be achieved with a query template with several parameters.
Eg: To create optional filters, you can use this pattern example:

match (n)
where
(case {{“field 1”:enum:{“values”:[“any”, 1, 2, 3],”default”:”any”}}} when ‘any’ then true else n.field1 = {{“field 1”:enum:{“values”:[“any”, 1, 2, 3],”default”:”any”}}} end)
AND
(case {{“field 2”:enum:{“values”:[“any”, “abc”, “def”, “hij”],”default”:”any”}}} when ‘any’ then true else n.field2 = {{“field 2”:enum:{“values”:[“any”, “abc”, “def”, “hij”], “default”:”any”}}} end)
return n
limit 10

Question 19

How does the timeline feature need to be programmed i.e. do you need timestamps in both nodes and edges?

Answer 19

You can have timestamps on both. No restrictions. See here: link

Question 20

Can we programmatically style nodes and edges (e.g. certain nodes or edges appears larger based on data)?

Answer 20

You can use styles: link. You don’t need to do anything programmatically as soon as they are set (default for datasource or by a single user) they will be applied automatically later on.
In any case you can use API to set the first configuration programmatically.