dedicated visualizer
Depth-First Search
Follows one branch as deeply as possible before backtracking to try the next branch. This page keeps the runner, chart, and controls focused on a single algorithm so the walkthrough feels calmer than the overview page.
session controls
Compare this algorithm against a related one, turn on quiz mode, or keep the current state in a shareable URL.
current shareable URL
Copy the URL to preserve this exact dataset, target, compare mode, and quiz state.
browse more
Want a different problem or visual mode? Jump back to the catalog and open another dedicated page.
open catalogscenario presets
Load a focused input that reveals a specific behavior quickly instead of hand-editing every value first.
graph controls
Graph algorithms reuse the same learning graph so you can compare traversal and shortest-path behavior side by side.
graph note
BFS and DFS emphasize traversal order. Dijkstra adds weighted relaxations and a cost-aware final route.
chart + counters
The visualization and the live counters stay together so each step is easier to read.
current action · push start
current action
push start
visited
0
final 5
stack
1
final 2
frontier
1
final 2
steps
1 / 13
stack
A
current node
none
run summary
Finished in 13 steps. G is the target node, so DFS stops here.
visited
5
stack
2
frontier
2
steps
13
final route
A → B → D → E → G
current explanation
Depth-first search starts by pushing A onto the stack.
simple explanation
DFS dives down one path at a time using a stack.
pseudocode
complexity card
best
O(V + E)
average
O(V + E)
worst
O(V + E)
space
O(V)
algorithm notes
intuition
DFS commits to one path until it cannot go any deeper, then backtracks.
tradeoffs
- Good for path existence and topological-style reasoning.
- Does not guarantee the shortest path.
- Can be implemented recursively or iteratively.
when to use it
Use when you want deep traversal behavior, backtracking, or graph structure analysis.
interview tips
- Clarify whether you mark nodes when pushing or when popping.
- Use DFS to discuss connected components and cycle detection.
what I learned building this
typed definitions
One algorithm schema now drives the catalog, counters, pseudocode, notes, and visual modes, which keeps the UI consistent as the lab grows.
replay over mutation
Precomputed steps made it much easier to synchronize explanations, metrics, quiz prompts, and scrubber playback without hidden state drifting out of sync.
portfolio framing
Shareable URL state, compare mode, and responsive layouts mattered as much as the algorithm logic because this page needs to teach clearly and still feel polished as a product.
more in this lane
Want a different take on the same problem family? These stay in the same category but change the strategy.