Ask HN: What low-code “dashboarding“ SaaS would you recommend in 2021?
I‘d like to aggregate the output of a few REST calls visually. What options do I have nowadays? Thanks! Grafana is pretty great. There are a lot of built in connectors, but its ability to query arbitrary REST, JSON, oData etc endpoints along with databases makes it super flexible for what you mentioned. https://grafana.com/grafana/plugins/simpod-json-datasource It has a great GUI for building queries, plus solid RBAC, great charting, and a very flexible alert system. Configure in their GUI, then get your dashboard's code in as JSON for deployment. You can also define dashboard in actual code, ie grafanalib in Python and others- which is great for DevOps. https://github.com/weaveworks/grafanalib They have nice stylesheets ready for TV displays, but it gets really fun when users get their hands on it and can zoom in and drill down on time-series data, filter with dynamic parameters and such. Grafana sucks in terms of UX, but it's free (it's main selling point) and supports many backends. Creating dashboards is so freaking painful. I have liked it for what it's worth. My only pain point is bulk editing (no way to add something to every graph on a dashboard) and copying graphs across dashboards didn't seem so easy last time I tried it. I wouldn't exactly call its charting abilities great, for example something as simple as aligning the Y axis over multiple charts can't be done, and for BI(-like) charts you need to go through a bunch of hoops to get them right. Also the alerting system is not flexible at all, it's primitive (but functional). All and all it's only through the flexibility of the database we use (Postgres) we're able to massage the data enough to display it the way we want. yeah Grafana is probably the default choice I see at most customer sites for dashboards! I can’t decipher their pricing at all. What kind of usage gets expensive? What pricing? Just get the open source version, install it on a local machine - or on any available server - and get going: You do not need to use the hosted offering in that case. Take a look at https://anvil.works - it's not just "low code", it's a full web-dev platform so simple that you can build a dashboard with a few lines of Python - front end, back end and drag-n-drop design. (I'm a founder.) The trouble with "low-code" is that you're stuck on rails - it's very easy to do a few things, but the moment you want something the creators didn't anticipate, prepare for a lot of pain. Anvil is the opposite approach - it's a dev environment that happens to be simple enough to build a dashboard in minutes, but with a Real Damn Programming Language. It's basically Delphi for the web: It's got a drag-n-drop designer, Python in the browser and on the server, code editor with full-stack autocomplete, etc. (And unlike "low-code" tools, the runtime engine is open source, so you're not locked in: https://github.com/anvil-works/anvil-runtime) I love the value proposition of full stack w/ nothing but Python here. Backend has been cool enough with Python but for anything serious in the frontend, js and friends is almost always a necessity. Bridging that gap seems pretty lucrative, especially with your python → js compiler.¹ Does the open source runtime has the same Free-Tier restrictions (Anvil Branding, Libraries, Domain)? Not at all! People still upgrade for editor features - eg we have several Business plan users who primarily deploy with the open-source App Server - but it's by no means required. This is an incredibly appealing proposition! Well done. Retool is great software and has a great team. (https://retool.com/) Can confirm that Retool is amazing. Ran into some limits one or two years ago when using it with Firestore and had to go with something else, but I can’t really blame it for having trouble dealing with relationships when using a non-relational database ^^’ Retool is amazing for building internal control panels and support tools. I've been using Retool for a couple weeks and it's pretty impressive. Nice! It’s like SAP, but for sane people. It is a platform lock-in. So one must have to do an ROI analysis whether it makes sense to go with such a deeply coupled tooling. Or you could look at it as a tool for quick and easy prototyping. damn thats pretty cool There are so many such solutions, that It depends on your use cases. If you're looking for high-level dashboarding/monitoring that connects directly with your sources (REST, 3rd-party apps like Facebook, Adwords, etc), then tools in the realm of Klipfolio, Databox, Grow.com will suit you well. If you're looking for visualization tools that connect to your data-warehouse/SQL database, then BI tools like Holistics, Metabase, Redash, etc will likely work for you. And of course there's tools that can support both kinds of sources, like PowerBI and Google Data Studio. P/s: This post does a good job at laying down a framework for evaluating these sort of BI tools: https://www.holistics.io/books/setup-analytics/navigating-th... This is a great reply. A nondev should likely stick to the first group, which has a lot of integrations and the hardest coding will be around excel formulas at most In my experience its best to do the following: Have 1 dashboard specific to business finance and ads metrics. This dashboard includes klipfolio or simpler things like grow or fathom accounting. Another tool connected to your bd, at the most basic starting with mode Analytics and then going full visual with looker or tableau (this is where things get expensive) You seem to be kind of shadowban (not sure what the proper term is), you might want to write an email to the hn moderation to clear that up you were correct. Thanks for pointing this out. Have a great new year! Grow.com supports both 3rd party, as well as SQL database. It doesn't run SQL directly on your db, but pulls it into the provided data warehouse. I work at Grow, happy to answer any questions. Send metrics to AWS Cloudwatch as custom Metrics. Then get a nice hefty bill. CW is expensive. You can give https://chartbrew.com/ a try if you wish. You can make the API requests in a Postman-like interface directly on the platform. So you can have live-data coming in as the tool makes the requests automatically for you. You can also have multiple requests made to different services gathering data on the same chart. Disclaimer: I'm the creator of the tool, so feel free to ask any questions about it and I'll be happy to answer I looked into this recently, when I started writing the software to analyze my selftracking data. After researching available options for a while and consulting everyone I know in this field, I settled on building on top of the Plotly Javascript library: https://plotly.com/javascript/ It does nice interactive charts out of the box. All you need to do is throw data at it. What it lacks are more high level operations like: - Interactively switch between different chart types (lines, scatter, box plots, bar charts etc) - Aggregate time series by hour/day/month/week/year - Extrapolate data to the rest of the current time period So I built this functionality on top of Plotly and I am happy with the result. It looks like over time I will turn this into a mix of data studio and jupyter notebooks. But it runs in the browser and is simple and blazingly fast. I picked plotly in 2019 to build kind of a low-code-dashboarding at my company and got to look into it again few weeks ago, noticing there in fact is an aggregation capability. I really like Trevor.io It has nice gui for data exploration, clean charts and integrates really well with gsheets. If you save the API calls in a database could use this? Appsmith would work great! It’s an open source project with cloud hosting and self-hosting options. You can query any REST endpoint and use JS to aggregate the outputs to connect it to charts, tables or text components. Also use JS to change the query params when you use date or search filters.
(I’m a founder). https://github.com/appsmithorg/appsmith
www.appsmith.com I'm currently working on my own little dashboard SaaS - https://statsout.com/ It doesn't have the functionality to perform custom REST calls yet but it's definitely on the list. For now you can get the basic metrics from Google Analytics, Search Console, Facebook Pages, Instagram, and Stripe. It's still very much in beta but you're welcome to try it. If your dashboarding needs are as rudimentary as you describe, then Streamlit! https://www.streamlit.io/ I use it for my dashboards. It's super simple to prototype apps and it's under rapid development. That, or if interactivity is not that much of a concern, Jupyter notebooks (maybe with Voilá) is also a good option. I'm currently working on a SaaS like this: https://www.rationalbi.com/ It's meant to be a modern take on BI/Analytics, and sort of a blend between dashboards, scheduled reporting and notebook analytics (like Observable/Jupyter). You can just drop in data files of any kind, syndicate data from your cloud drives, connect to remote databases, upload via the API, and it has 80k+ open datasets available for analysis and benchmarking. It also has a full SQL database running in the browser, so if you build dynamic notebooks/dashboards with business logic, you can do really fast SQL without any network latencies. Disclaimer: I'm the creator/founder of Rational BI. Had investigated Apache Superset recently but was apparently in a to early stage for production
https://superset.apache.org/ Definitely on my watch list though! We don't work with rest calls - just directly with databases and SQL, but do take a look at chartio.com! We aim to make it as fast and agile as possible to make and collaborate on charts and dashboards with your team. metabase.com is amazing if you're dumping the data into Postgres or similar. I want to love Metabase, I really do. But some of the UI decisions are mind boggling. Why is accessing reports and dashboards so difficult? Things are hidden away under poorly named constructs. Buttons don't do what I'd expect them to do... It's great but it comes with a few drawbacks. I didn't find this at all - it's great to be able to pin a bunch of dashboards to the main page that users see when logging in, and to allow them to drill down from the dashboard by filtering their own data, or downloading a CSV. The UI for creating dashboards isn't the best, but it's really nice to be able to drop down to SQL so easily and share or embed results. Their embedding is pretty darn decent. You could do something like set up a key or key per role in Wordpress and embed them as a post type (or your CMS of choice) in an Intranet if you wanted a little more control over it. You need to set up a signing key on the backend, which is a good thing. Many dashboard sharing schemes end up being public or with very weak security. I second this. I used database about 5/6 years ago and it was amazing. But they started to add collections a d my collections and a bunch of stuff and now everything seems hidden. I still use it (deployed it at my current company and the CEO loves h iui s dashboard). But I hate the way things are hidden in the current layout. I'm working on a tool to build any sort of dashboard by just a couple of lines of JS. This is our YC application video https://www.youtube.com/watch?v=0XfCwXCPLcU&feature=youtu.be We didn't get in (this time), but we are about to launch a beta soon. If someone is interested please send me an email at <username>@gmail.com - Thanks Really looking forward to seeing the progress of causal[0]. Looks to me like they have a shot to become the best in class for interactive "dashboarding" in 2021. Imagine being able to do some on the fly scenario planning working from live data. Could be a game changer. [0] - https://www.causal.app/ thanks a bunch! "on the fly scenario planning from live data" is exactly what we're trying to enable :) Grow.com is a great low-code / no-code option. Anyone can use the graphical transforms to 'write' SQL and you can pull all the data into the provided data warehouse so don't have to set that up. Is is not cheap, but it is powerful. I work there, happy to answer any questions. I recently evaluated a bunch and https://github.com/appsmithorg/appsmith covered our needs. For building internal tools seems like a great alternative. Glad that it works for you and thanks for posting about it here. (I'm one of the co-founders) Forest Admin, if we're talking about more than dashboards/charts and you need to dig record deep to perform CRUD operations - https://www.forestadmin.com I'd recommend MintData (https://mintdata.com) It lets you define your dashboard logic in a spreadsheet & then drag/drop the dashboard UI and publish it for end-user consumption. Check out Superset. https://github.com/apache/incubator-superset It’s modern, easy to extend. From the same author of apache airflow. We used SumoLogic at my last job, really nice for dashboards IMHO, much easier than Grafana, it's log oriented but works with any structured data you throw at it.
I think it was expensive though. If you want something more “hackable” (making changes on data, or use SSO), I recommend going with Abstra (https://abstra.app) Full disclosure, I made this tool What does the output look like? I've recently had good luck streaming stuff from python into InfluxDB 2.0 - and that has built in basic charting / dashboard. Not grafana level fancy, but not bad either If you thinking anything related to images, Try Bruzu.com Forrestadmin looks good also. Haven’t tried it though Sisense is pretty cool and you could connect up a bunch of different datasources Visual JavaScript is low code and has been used to build dashboards yazz.con We are working on a Blockchain SaaS (Zeeve - https://www.zeeve.io) to automate the complete DevOps and SecOps approach. Also to add up integrations with grafana, promethius etc for monitoring and analytic. Zapier/IFTTT with Google sheets & apis you want monitor? It'll be an unpopular answer but.. Google Data Studio...? - Grafana - Rails Blazer gem (if you are using rails) Bubble.io I’m also a fan of Bubble and have made a few dashboards and apps with it. Google Data Studio? I personally used Redash in the past, and found it very easy to create dashboards and even customize our instance. Not sure how it’s evolved though since being bought out by Databricks.
Visualize/Filter/Add Alerts using all the builtin tools from Amazon Cloudwatch, you will need no servers and no configurations. Just run in a cron job a set of queries every 5minutes aws cloudwatch put-metric-data
--metric-name registered_users
--namespace MySaaS
--value 241