Avis et détails du produit de Hex

Hex is a wonderful product. It is a souped Jupyter notebook with native SQL and some nifty dashboarding features. For people able to code is a far superior solution compared to point and click BI tools like Tableau and Power BI. Avis collecté par et hébergé sur G2.com.

The three things I feel most limited by are:

1. It would be really nice to be able to support Python packages like pymer4 that rely on an R instance to run. There are many statistical packages that are R only.

2. The R functionality is basic compared to what is supported in Python. Would liek to see that built out.

3. Hex is pretty limited computationally compared to even my local machine. I often run into a bottleneck in terms of pure computation. Avis collecté par et hébergé sur G2.com.

Hex is so helpful for visualizing data and translating complex data insights into easy to digest insights for non-technical stakeholders. I compare Hex to Looker a lot and I love the ability to dynamically construct queries to create visualizations that improve performance and run times.

I use Hex every single day and create 3-4 dashboards every 2 weeks for stakeholders. The Hex apps have been widely accepted by stakeholdes. Avis collecté par et hébergé sur G2.com.

Hex does have some limitations. I recently came across a need to separate permissions for different users within a single app between tabs. I believe the current workaround is to embed individual cells into a separate application for visualization but if Hex could allow for different user access within a single app, that would be really great! It is kind of a niche use case but would definitely be appreciated. Avis collecté par et hébergé sur G2.com.

As a new user, I am still getting familiar with HEX and all of its features/functions. I like the sharing capabilites, it really assists with seamless collaboration. I appreciate the new outline feature to help me stay organized. Avis collecté par et hébergé sur G2.com.

For me, the visualizations are a bit of a downside. I find myself constantly having to rewrite queries or use parameters (work arounds) to make simple visualizations or big number values that are really easy to do in Mode or Tableau. For instance, why not be able to add more series to y axis for grouped charts. For example, if I have 3 values I'm trying to show over time, my only option is to do one as a bar, one as a line, and can't do the other at all. And this is just for the left axis. If I want to do the combo, then I would have 2 lines and it wouldn't make sense and be confusing. Also, it would be helpful to add filters to visualizations where end users can't see them. Avis collecté par et hébergé sur G2.com.

J'aime le cluster toujours connecté et l'exécution rapide du code dans HEX, contrairement à Databricks où je dois toujours connecter les notebooks au cluster pour exécuter du code. J'ai interagi avec le support client une fois et ils étaient vraiment serviables et facilement disponibles pour aider via Slack. Je trouve le codage et la mise en œuvre d'une solution à la fois en SQL et en Python faciles sur HEX sans trop de tracas. Je suis data scientist chez Upwork et j'utilise HEX presque quotidiennement pour travailler sur des problèmes de science des données en utilisant Python et SQL. Je trouve l'intégration des notebooks HEX avec d'autres plateformes de stockage comme le bucket S3 simple. Avis collecté par et hébergé sur G2.com.

Je regrette la facilité de déployer n'importe quel carnet sous forme de pipeline et de le programmer sous forme de flux de travail dans HEX. Avis collecté par et hébergé sur G2.com.

It's very clear that the developers and product team at Hex have interviewed a ton of data professionals and are making a really great tool. The recent addition of Sections is proof of that, the amount I had to scroll through my longer notebooks was 100% my biggest complaint. Compoments and other features like that add a ton of flexibility to hex, and the customer support has always been helpful whenever I've needed it. Avis collecté par et hébergé sur G2.com.

I still struggle with extremely large datasets in hex - being able to return sql cells as queries and other workarounds like that certainly help, but when returning the data as a pandas dataframe things get bogged down very quickly once there are more than 1 mil rows in play. My other complaint is that the charts are not the easiest to use - if I want to very quickly visualize a bunch of different slices of data I will frequently turn to another tool and then return to hex once that step of my process is done. Other than that there are just weird little bugs sometimes, like the order by clause in a sql statement not reliably returning results in the requested order. Avis collecté par et hébergé sur G2.com.

The notebook experience in Hex is unmatched when compared to other data visualization tools in same class (e.g. Mode, Tableau, and Hex). This is especially true for users that are comfortable with code e.g. Python, SQL, and R. Additionally, Hex integrates well with the most common data warehouse types e.g. Snowflake, Clickhouse etc. The native charts are also very helpful and generally Hex appears to prioritize support for visualization libraries in Python which helps to cover use cases outside of what is available in native Hex charts. Avis collecté par et hébergé sur G2.com.

The only item that keeps me from giving Hex a perfect 10 score is that my experience with customer support has never been really been great. I'm not exactly sure why, but Hex customer support always takes a while to respond and when I do get a response it typically isn't very helpful. I suspect that the customer support team doesn't have great visbility into the issues facing customers, but I am unsure.

That being said, other less significant issues I've run into are:

1) Attempting to use GPUs on Hex. My team tried to use a GPU instance but frankly was unable to due to issues with package / library management. Honestly though, GPU instances feel a bit odd on Hex to begin with, and there are many other service providers I would rather use for development on GPUs.

2) Hex Magic does not work well at all. I personally don't have a strong need for an embedded LLM / AI into my Hex notebooks, but any time I've actually tried to the Hex Magic feature it often underperforms when compared to base ChatGPT. This is coming from someone who uses LLM / AI regularly and also develops applications that leverage LLM / AI. I suspect something is off with the underlying evaluation approach / pipeline. A good Hex Magic experience would feel like using Cursor, where the model understands the full context of the project. Avis collecté par et hébergé sur G2.com.

J'aime que Hex offre la possibilité d'intégrer de manière transparente des données de Snowflake dans un IDE semblable à un notebook Jupyter, ce qui me permet de créer rapidement des applications de démonstration à partager avec d'autres personnes qui sont toujours à jour avec notre base de données. De plus, la possibilité de programmer des exécutions est très pratique. Avis collecté par et hébergé sur G2.com.

Importer des packages depuis notre propre dépôt GitHub privé et récupérer des packages depuis pip est bien, mais ce n'est pas toujours aussi fluide que je le souhaiterais. De plus, les packages disponibles par défaut ne sont pas toujours à jour (tiktoken, par exemple). Ce serait bien s'il y avait un autre moyen de spécifier de manière fluide quels packages et quelles versions de packages vous souhaitez inclure dans votre application. Avis collecté par et hébergé sur G2.com.

I absolutely love how I can query from many sources -- my job typically has large CSVs that are unmanageable in Excel / GSheets, and I love how I can query from both our Snowflake source AND these CSVs. It's been a game changer. I can write complex code and have a digestible format for someone at the end. Avis collecté par et hébergé sur G2.com.

I've found that some of the visualization options are a bit limited (can't create totals on a pivot table) Avis collecté par et hébergé sur G2.com.

Quite easy to navigate, filter, and create queries. In general fast and easy way to search in data base for not engineers. Avis collecté par et hébergé sur G2.com.

One little thing. When searchin in table mode it's not possible to easy download results, as it's for code/query. I need to click "table result" which is under "table", and then download. Now I know that but initilly I thought that it's not possible to download results when I search in "table" mode. Avis collecté par et hébergé sur G2.com.

Hex ne cesse de me surprendre. Cela fonctionne tout simplement. C'est un outil puissant qui est également simple à utiliser. Chaque fois que je me demande si je peux utiliser Hex pour résoudre mon problème, la réponse est presque toujours positive. J'adore leurs outils de visualisation de données, la rapidité du produit et leur assistant IA. Avis collecté par et hébergé sur G2.com.

Pour être honnête, je n'en ai trouvé aucun. Quelques fois, j'ai pensé que ce serait génial si Hex avait des fonctionnalités xyz, et peu de temps après, l'équipe les a ajoutées. Avis collecté par et hébergé sur G2.com.